[0:00] WHAT IS UP EVERYONE AND WELCOME to the W
[0:04] show. We've got a great show lined up
[0:06] for you guys today. Google has been
[0:08] found liable for false information
[0:12] provided by AI overview. This is in
[0:16] Germany, so it might not necessarily be
[0:18] helpful to everyone, but it is a step
[0:21] towards holding big tech accountable for
[0:25] the false information that it propagates
[0:28] with its um, you know, let's call them
[0:31] sophisticated and yet foundationally
[0:35] flawed AI technology. In other news,
[0:38] speaking of flawed technology, Donut
[0:41] Labs's solidate battery has been exposed
[0:45] as a regular lithium ion thanks to some
[0:49] incredible work from one of our our
[0:52] fellow YouTubers who I Oh man, I hate it
[0:56] when people do this. Uh, so whose name
[0:59] I'm actually going to use, unlike the
[1:01] way that people usually cite YouTubers,
[1:03] Zeroth, fantastic video we're going to
[1:05] be talking about a little bit later. No,
[1:07] don't you hate that, Luke? When the
[1:08] headline of a traditional media article
[1:10] is YouTuber,
[1:12] >> yeah,
[1:13] >> YouTuber has a name. Uh, what else we
[1:16] got this week?
[1:18] >> You can just tell the Instagram
[1:20] algorithm what you want now, which is
[1:22] which is actually kind of interesting.
[1:25] And this is a this is a sounds really
[1:28] fantastic on the surface and is kind of
[1:31] but is also terrifying. Uh a Florida
[1:35] hospital is using Palunteer
[1:37] to catch sepsis earlier and it's saved
[1:40] 886 lives so far.
[1:44] So is this one of those cases where like
[1:47] even a broken clock is right twice a day
[1:49] Palunteer did something that we're
[1:52] supportive of?
[1:54] Uh well
[1:56] let's get more into it.
[2:19] The show is brought to you today by ODU
[2:21] Op manager Nexus, Squarespace, and
[2:24] Server Parts Deals alongside Dbrand, our
[2:27] rap partner. Razer, our laptop partner,
[2:29] and also Razer, our oops, that logo is
[2:32] supposed to be visible, I think, our
[2:34] chair partner. I'll be getting to this a
[2:35] little bit later. This is super cool,
[2:37] but but it will come later. All right,
[2:39] why don't we jump right into our
[2:41] headline topic, which is that Google has
[2:44] been found liable for false information
[2:47] provided by AI overview. In a potential
[2:50] landmark case, the regional court of
[2:52] Munich has ruled that Google, if they're
[2:55] going to, you know, have their product,
[2:57] spit out an output based on all of the
[3:00] data that they ingested, some of it
[3:03] legally, some of it, it seems, uh, we're
[3:05] still questioning the legality of it.
[3:08] They're accountable for what it says,
[3:10] which, you know,
[3:12] maybe is a good thing. The same way that
[3:14] newspapers were accountable for what
[3:16] they printed and you know radio shows
[3:19] are accountable for what they say. Well,
[3:21] Google's AI is also accountable. Yeah.
[3:24] PE, if you're a if you're a mass media
[3:26] outlet and you reach literally millions
[3:28] of people, you should have some
[3:30] accountability for having there be a
[3:33] factual basis to the things that are
[3:34] being propagated.
[3:36] >> Sure. I just don't know how often that
[3:37] actually like applies. It it applied to
[3:40] um uh what's his name? Um,
[3:47] >> you're not going to bring up Alex. You
[3:48] are okay. That's where you're going with
[3:50] it. Uh, well, yeah. I mean, it relies on
[3:52] somebody to actually hold you
[3:55] accountable to it in many cases, but
[3:57] there is there is a basis for but there
[4:00] is an expectation that if you are
[4:03] spreading information that you are
[4:06] responsible for the accuracy of it. The
[4:08] case was initially brought by two German
[4:11] publishing companies after Google's AI
[4:14] overview presented information about
[4:15] them, combined with that of shady
[4:19] companies and online scams, and then
[4:22] drew information that didn't appear in
[4:24] any of the linked sources,
[4:28] which I mean, Luke, have you ever
[4:29] experienced that with Google's AI
[4:31] overview where it goes blah blah blah
[4:33] this thing citation and then you click
[4:35] it?
[4:38] completely not in there. Like
[4:40] >> citations by AIS are are are pretty
[4:42] sketchy in general, but um this this
[4:45] does line up. This was pointed out by
[4:47] Flowplane chat as well, but I I think I
[4:49] lost the message, but this does line up
[4:51] with other things. There's uh customer
[4:53] support chat bots where what they say
[4:56] has been held as like fact. So the
[5:00] customer support chatbot offers you
[5:01] something and uh the the company's held
[5:05] to to follow through with it. Um, did I
[5:08] did I just see it? I thought I just saw
[5:10] it. Darn it. Somebody in chat pointed
[5:11] out I think it was Air Canada. Yeah, Pa
[5:13] Paige in chat said Air Canada was
[5:15] already held liable for something
[5:16] incorrect their support chatbot said.
[5:18] And I think there was other companies as
[5:20] well, not just Air Canada. So,
[5:23] >> Google,
[5:23] >> yeah, it kind of follows other things.
[5:25] Google pulled the classic defense of,
[5:27] "Well, users can fact check for
[5:30] themselves, and they do plan to appeal
[5:32] the ruling with a spokesperson saying,
[5:34] "This case focuses on specific and
[5:36] narrow errors, not the foundational way
[5:39] that AI overviews display web content."
[5:43] Um,
[5:45] right. But
[5:47] these specific and narrow errors are
[5:51] inherent to the way that foundationally
[5:55] AI overviews seem to work, at least at
[5:59] this stage in the game. Like these
[6:00] hallucinations are not
[6:03] new. They're not an unknown. Um, and to
[6:09] Google's credit, they do say in AI
[6:12] overviews that, you know, AI can make
[6:14] errors and that you need to check it or
[6:16] whatever, but like
[6:19] I I think there's a
[6:24] I think there's still a if we all know,
[6:28] right? It's kind of like the it's like
[6:30] what went down with end user license
[6:32] agreements, right? like you had these
[6:34] these long theoretically legally binding
[6:37] documents.
[6:39] Everybody knew on both sides, the
[6:41] companies writing these end user license
[6:43] agreements, the people signing them,
[6:45] everybody knew that nobody was reading
[6:47] them. So when push came to shove, it was
[6:50] ultimately found that if if someone was
[6:54] not actually wellinformed and was was
[6:57] not actually aware, then this stuff
[7:00] wasn't going to be binding. it didn't it
[7:02] didn't just absolve you of of breaking a
[7:05] law the fact that you wrote into your
[7:06] end user license agreement that you were
[7:08] like allowed to do it. So, in the same
[7:10] way, if we know that people are trusting
[7:12] this stuff and that they're not noticing
[7:14] these these errors and they're not
[7:16] actually doing the due diligence, then
[7:18] there, in my opinion, there's a certain
[7:20] responsibility on the companies that are
[7:22] putting this information. I'm going to
[7:24] put that in giant air quotes, putting
[7:26] this information out there to ensure
[7:28] that it is factual. And if it's not
[7:30] factual, who's accountable? Who's
[7:33] responsible?
[7:35] And yeah, Google's going to say the
[7:37] user, but like I think Google knows
[7:41] probably better than just about anyone
[7:43] else that the user is just not going to
[7:46] do that work. Like the entire the entire
[7:50] premise of this feature is that we're
[7:53] just going to surface the information in
[7:55] an even more convenient manner because
[7:57] we know the user doesn't want to click
[7:59] through and read a bunch of stuff.
[8:01] That's the whole point.
[8:03] >> Yeah. Yeah. So, we have a few discussion
[8:06] questions here. Um, first of all,
[8:10] should the AI pushers be allowed to use
[8:12] the public as their beta testers while
[8:14] they work out the kinks here? Is there a
[8:16] level I I'm going to clarify this or I'm
[8:18] going to add a layer to this one. Is
[8:20] there a level of disclaimer that you
[8:23] would consider
[8:25] acceptable so that it could just spew
[8:27] misinformation like this? Like, how beta
[8:30] would it have to look and feel?
[8:34] Um, I'm trying to see what it looks like
[8:35] right now, just out of curiosity. Sure.
[8:38] Um, it doesn't look like there's
[8:42] any flags whatsoever. So, above the
[8:45] fold, um, I I just Googled testing and
[8:50] it gave me an AI overview of the word
[8:52] testing. Um, and before clicking show
[8:56] more, there's there's nothing. It it has
[8:58] the little Gemini icon. It says AI
[9:01] overview. There's the the quick summary
[9:03] with its uh citations,
[9:06] some more information, more links on the
[9:08] right hand side, but there's no
[9:09] indication that this isn't ready to go.
[9:12] >> Yeah, I've got I did the same thing just
[9:14] to to make it a little bit easier for
[9:15] the people. So, it's got a couple
[9:17] citations here that makes it look very
[9:20] credible.
[9:21] >> I click show more.
[9:22] >> Once you click show more, there's fine
[9:24] print at the very bottom that says AI
[9:27] can make mistakes, so double check
[9:29] responses.
[9:29] >> Holy crap. I I actually do do I have
[9:32] that? Am I missing it, guys? Am I
[9:34] missing something here?
[9:36] >> It's on mine. It's right above where the
[9:39] chat input is where you would type just
[9:42] to the left of the like share, like, and
[9:44] dislike buttons.
[9:45] >> Uh-huh.
[9:47] Am I Am I Guys, help me out here. I feel
[9:50] like I'm gaslighting myself because I
[9:53] was pretty sure there was a thing. I do
[9:56] not see it.
[9:56] >> Whoa. I don't see it on yours. What the
[9:59] heck?
[10:00] >> Mine has it.
[10:02] What?
[10:04] What the heck?
[10:07] >> That's super weird. It's It's on my
[10:10] side.
[10:11] >> Okay, doesn't always show.
[10:13] >> People are asking if I have an ad
[10:14] blocker. Uh, I mean, it shouldn't do
[10:17] that, but like, what extensions do I
[10:19] have?
[10:20] >> I got nothing here, guys. Here's my
[10:21] extensions.
[10:23] What What else we got here, guys? What
[10:26] else we got?
[10:28] Do you want to open a a private browsing
[10:30] tab and do the same thing?
[10:31] >> I would like to do that very much.
[10:37] Okay.
[10:40] Blah blah. There it is.
[10:43] >> Yeah.
[10:44] >> Whoa, whoa, whoa, whoa. Okay. So, that's
[10:45] interesting. It also moved this
[10:47] interface over to the right.
[10:50] Yeah, it's probably just reactive, but
[10:52] yeah,
[10:54] >> it didn't need
[10:57] Yeah, but it's It went all the way from
[10:58] left aligned to right aligned, though.
[11:00] Oh, no. It's not quite left aligned. Oh,
[11:02] you're right. Interesting.
[11:05] Is it even quite right aligned, though?
[11:07] You know what? No, no, I think
[11:08] >> Yeah. No, it's just buffered off on both
[11:10] of them.
[11:10] >> Interesting.
[11:12] So, it just it just didn't for some
[11:16] reason sometimes. I mean man.
[11:20] Okay, so back to my question then. What
[11:23] is the degree of disclaimer that would
[11:26] be acceptable to you? Obviously what I
[11:29] encountered just there with no
[11:32] disclaimer and this is this is Chrome.
[11:34] This is like the chromest chrome ass
[11:39] Chrome of Chrome browser. I don't even
[11:40] have any extensions or anything like
[11:42] there is no excuse because I'm Googled
[11:46] all the way through the pipeline. So
[11:47] clearly that's not acceptable at all.
[11:50] And honestly, I don't even think the
[11:52] other version that I got in the
[11:53] incognito window, I mean, I think you
[11:55] and I could probably both agree that's
[11:56] not enough if it's going to be making
[11:59] like egregious errors like this.
[12:01] >> Yeah, fine print that you have to click
[12:04] show more to even see. Like it's
[12:06] literally not like you you can get
[12:09] results
[12:11] that are conclusive results and not see
[12:15] the disclaimer because you didn't click
[12:16] show more. You need to click another
[12:19] thing in order to get the disclaimer is
[12:20] like obviously not enough. I think it
[12:22] would need a pretty major badge. Um I
[12:25] don't look at the AI overview too often
[12:27] so I wasn't really ready for this but I
[12:29] was expecting it to next to where it
[12:31] shows the Gemini logo and AI overview. I
[12:34] was expecting a big fat beta testing
[12:37] batch. Um,
[12:41] >> right. So, you want it like right here
[12:44] like AI overview
[12:46] beta uh you know results.
[12:49] >> Yeah.
[12:50] >> May contain uh misin factual errors and
[12:54] uh incorrect information. Uh check all
[12:56] cited links in order to uh verify before
[13:00] believing anything it says. That would
[13:01] be would that be enough for you?
[13:05] Uh, I mean, I don't like the feature at
[13:07] all, but yeah, I think that would be
[13:08] enough for me. Um, someone in chat, uh,
[13:13] Mini Catalyst said, "Dumb user here. I
[13:15] didn't know it was in beta." I don't
[13:18] think that it is.
[13:21] >> Um,
[13:21] >> I don't think it's in I haven't seen I
[13:24] don't see any indication on the feature
[13:26] that it's in beta.
[13:28] Yeah, I just said beta uh because
[13:34] I don't know
[13:37] that's the type of thing that I'm used
[13:39] to flagging, I guess. Uh right, speaking
[13:41] for them. Um, so yeah, like uh I guess
[13:45] don't flag it with beta, but
[13:48] some much more in-your-face thing saying
[13:51] that this information
[13:53] um is man, I I'd like it to even say
[13:56] likely incorrect because I catch the the
[13:59] very few times that I've really used the
[14:02] AI overview on Google, it's it's wrong
[14:05] like
[14:07] >> a lot of the time. I don't know what
[14:08] model they're running, but it seems not
[14:10] sophisticated. It's It's fast. Like it
[14:13] it really does seem like they're running
[14:15] a pretty fast, pretty light,
[14:17] >> pretty
[14:19] not thorough
[14:21] model. And the funny thing about it is I
[14:24] forget what I was looking at earlier
[14:25] today. Um, but I was I was sitting in I
[14:28] was sitting in script review with Elijah
[14:31] and
[14:33] for
[14:34] and maybe this is maybe this is crazy,
[14:37] but we uh Oh, yeah. I remember what it
[14:39] was. Um, we were looking for comments
[14:42] like you know how sometimes in our
[14:43] videos we'll do like a collage of
[14:46] similar comments like last time on tech
[14:49] house you guys asked you know why we
[14:52] scavenged the copper the way we did and
[14:54] we'll go d and we'll put a bunch of
[14:56] comments about that. So, there's a
[14:58] really cool new feature in the YouTube
[15:00] dashboard where you can search your
[15:02] comments using AI search for anything
[15:05] that instead of just searching for like
[15:07] a keyword, you can look for anything
[15:08] about this topic or anything that was
[15:12] critical of Lionus' hairdo or whatever.
[15:14] You can just you can search for for
[15:16] sentiment now, which honestly
[15:20] pretty cool and useful
[15:24] uh feature for for LLMs. Anyway, um, so
[15:27] I searched for something specific. I
[15:29] don't remember exactly what it was, and
[15:31] it crapped out a list of all the
[15:33] comments, and I was like, "Oh, that's
[15:35] super great." And okay, here, Elijah,
[15:37] when you're going through and you're
[15:38] doing the guidance for this video, here,
[15:40] let me just copy this link to the studio
[15:42] dashboard to the search result, and I
[15:44] will throw that into the script um like
[15:47] like guidance column.
[15:50] And then I was like, wait, is that going
[15:51] to work,
[15:53] right? like is it I know it'll bring up
[15:55] the dash. I know I can link someone who
[15:57] also has access to the dashboard and I
[15:59] can even link them to the specific page,
[16:01] but is it going to is it going to link
[16:02] that search query?
[16:05] And so I clicked it to check it
[16:08] and it did.
[16:10] But I was like, hey, wait a second.
[16:13] I think it generated that result again.
[16:16] It didn't cache it at all.
[16:20] which honestly for something like the
[16:22] YouTube dashboard wouldn't surprise me
[16:24] that much. But for something like a
[16:28] super commonly searched term like this,
[16:32] like do did you get exactly the same
[16:34] result as me or did you get an LLM
[16:37] generated output fresh for you? Do you
[16:41] want to have a quick look?
[16:45] >> Um, yeah. So, mine I think I closed that
[16:48] tab already. Okay. I mean, you could
[16:50] just search again. I'm I'll read mine
[16:51] and you tell me when it stops matching.
[16:53] >> Testing is the systematic process of
[16:56] evaluating a system.
[16:57] >> Already off.
[16:58] >> Okay. So, here's a question for me and
[17:03] and you know I I am not a systems
[17:05] architect. I am not a deep expert on
[17:09] LLMs or anything of the sort.
[17:12] But what I do know is that right now we
[17:15] have big tech absolutely desperate to
[17:19] build more data centers in order to
[17:21] support their AI endeavors. Um we have
[17:26] just like a like a massive volume of use
[17:30] of products like search and we have this
[17:33] technology that we've had for you know
[17:36] all of computing um caching that
[17:41] it seems to me might be useful. So why
[17:44] would they run like a crappier model
[17:47] that to your point, yeah, I've noticed
[17:48] that AI overviews makes a ton of
[17:50] mistakes, like more than other better
[17:54] LLMs makes. So in order to manage their
[17:57] resource usage, wouldn't you think they
[17:58] would just like
[18:01] cache this or like cache common queries,
[18:05] update them once a day, once every
[18:07] couple days? like the the the definition
[18:09] of testing
[18:11] >> is not going to change between now and
[18:13] tomorrow.
[18:15] >> Yeah. And and I find it interesting
[18:17] because before if I would Google
[18:20] something like a a specific word and I
[18:23] would do this very often.
[18:24] >> Yeah.
[18:25] >> It would it would surface, you know, a
[18:28] definition. Um,
[18:31] sometimes I will I I've done this in the
[18:34] past and you know, maybe they don't want
[18:35] me to do this, but I've done this in the
[18:36] past where sometimes I'll Google
[18:37] something for a spell check.
[18:39] >> Yeah.
[18:39] >> Uh, because the Google search will be
[18:41] like, "Ah, you spelled this word wrong."
[18:42] Right. Um, now it'll think I'm like
[18:45] trying to talk to it
[18:48] and I run into some weird some weird
[18:50] scenarios. Um,
[18:54] yeah. I I like Google is genuinely
[18:58] significantly less useful for me and
[19:00] I've been using it way less.
[19:03] >> This is interesting.
[19:04] >> AO overview became a thing.
[19:05] >> This is interesting. They're doing at
[19:07] least a little bit of caching. Chat
[19:08] pointed out that my two search results,
[19:11] one in my regular browser and one in my
[19:14] incognito tab, did return identical
[19:17] results.
[19:19] H
[19:23] Okay.
[19:24] >> Interesting. Uh, try a different browser
[19:27] entirely.
[19:28] >> Uh, okay. All right. I'll open it.
[19:30] >> All right. Microsoft, are you are you a
[19:32] Microsoft representative?
[19:35] >> Are you trying to
[19:37] >> No. Well, I mean, you could use Firefox.
[19:40] >> I don't have it on here.
[19:42] >> You could use Chrome to get Firefox.
[19:45] >> I did get the same result.
[19:49] interesting
[19:50] >> because I know there was this thing
[19:52] recently, I don't know if we talked
[19:53] about it on my show or not, but uh
[19:55] Google Chrome is reserving like a bunch
[19:57] of space on your on your drive uh for AI
[20:00] stuff and I don't remember if that in
[20:03] included things like results and
[20:04] whatnot.
[20:05] >> Okay.
[20:06] >> Um so a different browser getting the
[20:08] same one is is interesting. Someone in
[20:11] chat Oh, I just saw it. Uh, Kuro Setsuna
[20:16] 29 said, "I wonder if it's region cache.
[20:18] I'm in a different region."
[20:19] >> Yeah,
[20:19] >> that would be interesting.
[20:20] >> Yeah, that could be it.
[20:22] >> Um, cuz mine goes, "Testing is the act
[20:24] of subjecting a system person or process
[20:27] to experiments." Uh, which is different
[20:29] than yours. Yeah, it's close but
[20:32] different right?
[20:34] >> Yeah. Yeah. A little bit little bit
[20:37] different.
[20:38] >> And I I refreshed the page and got the
[20:40] same thing. Let Let me try in a
[20:42] different browser.
[20:43] >> Okay. All right, this has become kind of
[20:44] an interesting rabbit hole. Sorry, guys.
[20:46] I know you were expecting Wan show, but
[20:49] uh
[20:50] >> I mean this is W show.
[20:51] >> This is kind of Wan show.
[20:53] >> Yeah.
[20:55] >> Testing.
[21:01] Oh,
[21:07] in okay
[21:10] in Firefox
[21:12] >> paste
[21:13] >> on this device. Um,
[21:19] I'm not logged in.
[21:21] >> Mhm.
[21:23] >> So, it's figuring out my geo location
[21:27] and making everything
[21:31] Uh, Mandrin, I'm guessing.
[21:34] >> Oh, that makes sense. Actually,
[21:37] >> it did not give me an AI overview at
[21:39] all.
[21:40] >> Okay, then. Yeah, that checks out.
[21:43] >> English
[21:44] trying to do URL things
[21:48] testing.
[21:49] >> Avon Fox says it's probably quicker to
[21:51] regenerate than to pull from cache.
[21:52] Yeah, but that's exactly what I was kind
[21:54] of wondering is like are they better off
[21:56] using a better model? Um, and then using
[22:01] caching versus a super light model that
[22:04] like makes a ton of mistakes. Um,
[22:06] especially when it's something that, you
[22:08] know, is a is a pretty static uh outcome
[22:11] like a like something that is not going
[22:13] to to change from day to day like the
[22:17] definition of a common word for
[22:18] instance.
[22:24] One second here. Sorry.
[22:27] There we go. Uh, can I change this?
[22:31] >> It's also interesting to me that it um
[22:35] it cites Cambridge dictionary
[22:39] >> as its first one uh with
[22:43] club.mministry of testing, whatever the
[22:45] devil that is, and also some YouTube
[22:46] short. Um, and then when I actually go
[22:49] there, oh, there's an ad. Oh, wow.
[22:52] That's a lot of LT store ads. Oh my
[22:54] goodness.
[22:58] Uh anyway, nowhere on the Cambridge site
[23:04] is like exactly that definition. And
[23:07] you'd think it would be it would be
[23:09] simpler and less errorprone to just give
[23:12] me
[23:13] >> That's a lot of LT store ads.
[23:14] >> That's a lot of LT store ads. Hey,
[23:16] >> holy crap.
[23:17] >> All the LT.
[23:19] >> How much money was just spent spamming
[23:21] your browser with that?
[23:24] >> I shudder to think. you know, it might
[23:26] have been pretty effective because it
[23:27] ended up on W show though. Um, okay. So,
[23:30] mine mine did not match.
[23:33] Um, but I'm wondering if So, so I went
[23:36] to specifically a like Canadian
[23:40] Google search thing. Um, I didn't
[23:43] actually realize this was even a thing
[23:45] at all. I I searched up Google Canada
[23:49] and uh it gave me like
[23:55] I don't know if you go to Google right
[23:57] now, does it show a like red and white
[24:00] Google logo with the Canadian Soccer
[24:03] Association logo and stuff flicking by?
[24:07] >> Uh
[24:08] sorry, this Yeah,
[24:11] >> let me see.
[24:13] I gotta wait for this thing to catch up.
[24:16] Yeah. Okay. Okay. So, I got that
[24:18] basically um and I searched on there.
[24:21] >> Yeah.
[24:22] >> And now I got testing is the act of
[24:25] evaluating, measuring or examining the
[24:27] capability, safety or characteristics of
[24:29] something and then blah blah blah blah
[24:30] blah blah. Is that the same as yours?
[24:33] >> Um I don't
[24:35] >> because that is very different than my
[24:36] previous one.
[24:39] >> Not quite. systematic process of
[24:42] evaluating a system, product or idea to
[24:43] determine its functionality, performance
[24:45] or safety. Yeah. So, it just it shuffles
[24:47] things around a little bit. And when you
[24:49] shuffle things a little bit, well, then
[24:52] >> you introduce the potential for errors.
[24:54] And it's just seems to be a feature
[24:57] rather than a bug. And that's just kind
[24:59] of what we're stuck with. And somebody's
[25:02] somebody's got to be accountable for it.
[25:04] And you know, Google's basically saying,
[25:06] "Hey, well, it should be the user." But
[25:08] I mean, we just with this the simplest
[25:10] of tests, the searching the word testing
[25:14] found that it will sort of move things
[25:16] around. It'll site sources that don't
[25:19] actually have exactly that definition.
[25:22] I just I and it's and it's hard because
[25:25] it's really tempting cuz it's right
[25:28] there. Like I find myself searching for
[25:30] something and going,
[25:32] "Right, right, right, right, right. I
[25:34] can't trust that. Don't look at that.
[25:36] look under that. Go under the fold and
[25:39] start clicking into things because it's
[25:41] just it's just so convenient.
[25:45] Why don't we jump into our next big
[25:46] topic? You want to pick one?
[25:49] >> Sure. Um let's let's tackle this because
[25:52] I'm interested in how this is going to
[25:54] go. Um
[25:58] a Florida hospital is using Palunteer to
[26:00] catch sepsis earlier. It saved 886 lives
[26:04] so far. Um
[26:07] yeah, Tampa General Hospital in Florida
[26:10] has used Palunteer built has used a
[26:12] Palunteer built system called the sepsis
[26:15] hub.
[26:16] >> Okay, that's a scary name.
[26:18] >> It is to save an estimated 886 lives
[26:21] since 2022
[26:23] having sepsis related deaths. The
[26:26] software pulls realtime data from
[26:27] electronic health records, lab results,
[26:30] clinician notes, and bedside monitors
[26:32] across roughly a thousand patients at
[26:34] once, watching for subtle early signals
[26:37] of sepsis that can get lost on a busy
[26:39] floor. Then alerts a rapid response team
[26:42] so flagged patients get antibiotics
[26:45] within an hour. It's also cut sepsis
[26:47] patient hospital stays by 30%. which is
[26:51] uh going to benefit all other problems
[26:54] basically. Uh context why this matters,
[26:57] sepsis kills around 350,000
[26:59] Americans a year, making it the leading
[27:02] cause of death in US hospitals. Tampa
[27:04] General isn't alone in its results
[27:06] either. A separate AI model called
[27:08] Composer tested by UC San Diego across
[27:11] two ERS showed a 17% drop in sepsis
[27:14] deaths after going live. Uh the
[27:17] discussion question is if an AI system
[27:19] could demonstraably have sepsis deaths
[27:22] deaths at one hospital, what's actually
[27:24] stopping every other hospital from
[27:26] deploying the same kind of tool? Um
[27:29] >> I think the thing stopping it is
[27:32] >> posture but also rigor.
[27:34] >> Um like they have to verify that this
[27:38] thing actually does work first. Um, I
[27:40] mean, now that it's working, I think
[27:43] that will probably help it to become
[27:45] something that would be
[27:46] >> you need to make sure it's not doing
[27:47] anything else potentially bad as well.
[27:50] Um,
[27:51] but it it is probably pretty good. And
[27:54] obviously, you can't exactly b 886
[27:57] lives. And I think uh the families of
[28:01] the people saved would probably even be
[28:05] annoyed that I even brought this up. But
[28:07] one thing that flagged me immediately
[28:09] when reading this is that Palunteer is
[28:11] receiving the health records, lab
[28:13] results, clinician notes, and all the
[28:15] details from the bedside monitors of
[28:17] every single patient at this entire
[28:19] hospital. Um,
[28:21] >> well, roughly a thousand patients. I
[28:23] don't know if it's every single
[28:24] >> I think that's everyone at the hospital.
[28:25] I think it's everyone at the hospital
[28:26] that has those things. Um, I could be
[28:30] wrong, but I I I I think that's the case
[28:33] because their their note here, um, the
[28:36] other thing that I I thought was
[28:37] interesting, um, is it's it's not just
[28:40] patients with sepsis
[28:42] >> because they're trying to do early
[28:44] warning detections for patients getting
[28:47] sepsis.
[28:48] >> Um, so I I think they're just if you're
[28:51] if you're that wired up, um,
[28:54] >> I think I think your initial assessment
[28:56] was right. Uh, by the way, this is
[28:58] hilarious. AI overview. Tampa General
[29:01] Hospital is licensed for 1354 total beds
[29:05] systemwide across a blah blah blah blah
[29:07] blah uh blah blah blah blah. About Tampa
[29:11] General Hospital. Tampa General is
[29:13] licensed for 982 beds
[29:16] right under it.
[29:18] >> Beautiful. Fantastic. I I I honestly I I
[29:21] I find mistakes in the AI overview like
[29:23] almost every single time I use it. It's
[29:25] I I hate it. It has genuinely made
[29:27] Google significantly worse for me. Um I
[29:31] I have seen myself using other tools
[29:34] because Google is worse for me now than
[29:37] it was in like I don't know
[29:42] 2015. Like genuinely 11 years ago. Like
[29:45] we've we've gone so far back that we're
[29:47] over a decade behind. Um, and you know,
[29:51] maybe maybe it'll eventually be better
[29:53] or something, but it's it's worse than
[29:55] useless right now. Um, but anyways, back
[29:59] back to the topic.
[30:01] Looking at Palanteer's playbook, um,
[30:05] they want to own all of the data in the
[30:07] world. They've they've basically
[30:08] publicly said that they don't want
[30:10] people to have any form of freedom and
[30:12] that people won't commit crimes anymore
[30:15] when there is genuinely no such thing as
[30:17] privacy because uh any any government in
[30:20] the world would instantaneously know if
[30:22] anyone did anything out of line. Um like
[30:25] they they are
[30:28] not good. Um, and I think it's going to
[30:31] be very very difficult for people to
[30:34] look at this thing that is making a
[30:41] massive positive impact. Um, 886 lives,
[30:46] having the deaths and reducing hospital
[30:49] stays for a common and major issue by
[30:53] 30% are not things that anyone can b at.
[30:57] Those are very serious numbers.
[31:00] >> Those are exciting things. I mean,
[31:01] anytime you talk
[31:02] >> difficult to deny,
[31:03] >> even to an AI skeptic, you know, one of
[31:06] the first things they'll say is like,
[31:08] >> "Yeah, when they when they announced
[31:11] this stuff, they promised it was going
[31:12] to revolutionize medicine. If that's
[31:14] what it was doing, then I would be fine
[31:16] with it." Like pretty much everyone I
[31:19] talk to who's like anti- AI is like okay
[31:23] but if it was this then it would be
[31:27] fine.
[31:28] It's just that unfortunately you've got
[31:31] companies like Palanteer that are doing
[31:34] that stuff where we're saving human
[31:37] lives which seems like a pretty useful
[31:39] and noble endeavor for AI but also using
[31:43] it as a means to collect that data to
[31:45] pursue
[31:47] a lot less comforting future outcomes.
[31:52] Um,
[31:56] >> yeah.
[31:57] >> So, I think that could be one of the
[31:59] things back to our discussion question,
[32:01] preventing other hospitals from
[32:03] deploying this same technology.
[32:06] >> I also think like you wouldn't really
[32:08] have a choice. Um, a lot of times
[32:10] depending on how bad your condition is,
[32:11] you just literally have to go to the
[32:13] closest hospital. Um, I've heard some
[32:17] weird stuff about the American medical
[32:19] system where like you're out of your
[32:22] insurance capture zone or something, so
[32:25] you have to get
[32:28] driven further something. I don't know.
[32:31] I'm a Canadian. Our healthcare system
[32:33] does not work that way. Um, but like
[32:37] this stuff's just going to be running.
[32:38] Like I I I doubt you're, you know,
[32:45] There's there's a lot of
[32:48] systems built around health care and end
[32:50] of life where like the funeral business
[32:53] is incredibly lucrative because
[32:56] when someone passes
[32:58] >> you don't want to see
[33:01] >> yeah people are very unlikely to say no
[33:03] >> y
[33:04] >> to upsells. Uh so the industry makes a
[33:07] lot of money. Uh the healthcare industry
[33:09] is very similar. it. There is a ton of
[33:13] money that goes into especially near end
[33:15] of life because if someone uh closely
[33:18] related to you is suffering, you are
[33:20] very unlikely to say no to um upgrades
[33:24] in in care.
[33:25] >> Um this is one of those situations where
[33:27] it's like are you okay with the health
[33:31] records, lab results, clinician notes,
[33:33] bedside monitor access, and probably
[33:35] tons of other things being mass data
[33:38] harvested from your loved one. Oh, also
[33:40] they might not die if you click this
[33:43] checkbox.
[33:46] A lot of people are going to click the
[33:47] check box. Yep.
[33:48] >> I I would wager pretty much everyone. Um
[33:52] but then that also does lead to
[33:56] the, you know, the super dystopian
[33:58] futures that specifically Palanteer has
[34:01] been saying they want to create. Ivonne
[34:03] keeps asking me what I want for my
[34:05] funeral, and I'm like,
[34:07] >> I'm dead.
[34:10] Do whatever makes you happy. I couldn't
[34:11] possibly care.
[34:13] >> Yeah,
[34:15] >> monetize it. I don't know.
[34:18] Charge admission.
[34:20] I'm sure there's at least someone who'd
[34:21] want to see me dead. They'll show up.
[34:26] Fair farewell. Farewell, inferior tech
[34:28] liners.
[34:30] >> Church admission. Oh my god. Um,
[34:34] >> I mean, you've heard you've heard me say
[34:35] that if you guys catch it on camera, you
[34:37] have to you have to monetize it, right?
[34:40] >> I have. I was actually I was talking to
[34:42] someone about that surprisingly
[34:44] recently, and I I uh
[34:49] Yeah,
[34:50] I I was like, you know, he said that
[34:52] before. I don't know if he still feels
[34:54] that way or not. So, I guess you do.
[34:56] >> Do it. I mean, how like we got
[34:58] everyone's got to get their severance
[34:59] and everything. if for whatever reason
[35:01] we can't keep going without me, right?
[35:02] You got to you got to milk it. Milk it
[35:04] as good as you can, as long as you can.
[35:06] And then when I'm when it's over, it's
[35:09] over. And it was it was a wonderful ride
[35:11] while it lasted,
[35:13] man. Okay. Sorry. Uh Katsune Lane in
[35:17] full chat said, "Luke's dedication to
[35:19] Lionus is a W show delivered live from
[35:21] beside his coffin." That would be so
[35:25] depressing
[35:26] because it would be so long.
[35:29] >> Open casket. I hope
[35:31] >> I'd have to sit there. I'd have to sit
[35:33] there like open casket for like 4 hours.
[35:36] Oh my god.
[35:40] Oh,
[35:40] >> here. Practice.
[35:46] >> I can't even. Hold on one second. You
[35:47] got to
[35:51] Oh,
[35:53] >> how long could we keep the streak alive?
[35:54] I mean, would you would you would you
[35:56] bring would you bring my ashes and put
[35:58] them put them next to you? You and you
[36:00] and Dan can just do the Wan show and
[36:02] just put put me on the shelf.
[36:05] >> I Yeah, I think
[36:06] >> from the N show.
[36:11] >> It's just the sponsor spots. the sponsor
[36:13] spots can just go to like a earn camera.
[36:19] >> We could uh we could make the lid go up
[36:21] and down like the Canadians on South
[36:22] Park.
[36:23] >> If anybody sends a if anybody sends a
[36:25] comment that's a direct question for me,
[36:27] you could just like you could just cut
[36:29] to my camera and
[36:32] then just we'll leave it there for a bit
[36:34] and then cut back and keep going.
[36:36] >> Awkward silence.
[36:39] The one thing, the one place I draw the
[36:42] line though is no LLM resurrection of
[36:45] me. I am one of the people on earth who
[36:49] would be relatively easy to to LLM in
[36:54] honestly a pretty accurate fashion,
[36:57] >> but I don't consent to that. I don't
[37:00] approve of that. Um, so so I'll just on
[37:03] the on the urn camera, I'll just have to
[37:06] like flap the lid of the urn to pretend
[37:08] you're talking and just
[37:09] >> Yeah.
[37:09] >> try to
[37:10] >> You have to puppeteer your voice.
[37:12] >> You can even do the voice. Okay.
[37:15] >> Yeah. Yeah. Yeah.
[37:16] >> I'm fine with that. I'm fine with that.
[37:18] >> I don't know who outside of Dbrand would
[37:20] accept those ad spots.
[37:23] >> I mean Dbrand would do it. If Dbrand's
[37:25] still around,
[37:26] >> they they would do it.
[37:29] Yeah,
[37:30] >> I can find out.
[37:32] >> Also, laughing about
[37:34] flapping your urn. I'm I'm in a rather
[37:37] public place. Uh just so viewers of the
[37:40] show know, I'm at um Lee Hua Hong. I
[37:44] I've I've featured this place a couple
[37:47] times. There was one time somewhere
[37:49] around 10 years ago in the desk and
[37:51] internet cafes video that I I featured
[37:53] this place. And then um there was also
[37:59] a short a little while ago, but I needed
[38:01] a good internet connection.
[38:03] >> Mhm.
[38:03] >> So I was like uh we were having some
[38:05] issues with hotels and their internet
[38:07] connections, so we just decided to do
[38:10] the show from here. Um and people can
[38:14] definitely hear me.
[38:15] >> Nice. Hold on. Hold on. Hold on. My
[38:19] rep's typing. There's a robot. There's a
[38:21] robot typing.
[38:23] I said you would sponsor wrapping my
[38:25] earn, right? Yeah, I'll schedule it,
[38:28] too. Classic.
[38:32] >> Yeah, not not surprised.
[38:35] >> Yeah, that checks out. I mean, they've
[38:37] tried to kill me enough times, sending
[38:39] me packages full of broken glass and
[38:41] whatnot.
[38:42] >> I got a comment.
[38:43] >> Maybe that's why they're just annoyed
[38:44] they haven't been able to do that type
[38:46] of an ad spot yet. They've been waiting
[38:48] for so long.
[38:51] >> Oh man.
[38:52] All right, we should probably jump into
[38:54] our next topic here. Uh, th this is on
[38:59] the surface kind of cool. You can
[39:02] apparently just tell your Instagram
[39:04] algorithm what you want. Now, Instagram
[39:06] is expanding its your algorithm feature
[39:09] to the main feed, letting you directly
[39:12] tell the app what you want to see. It
[39:14] shows you a list of topics Instagram
[39:16] thinks you're interested in based on
[39:18] your activity. And you can add topics
[39:20] that you want or remove ones that you
[39:22] don't. Changes will carry across real
[39:25] explore and now the main feed since it's
[39:28] all one system. It started in the US and
[39:30] is rolling out globally in English.
[39:33] Instagram head Adam Missouri framed this
[39:35] as fixing a one-sided relationship. The
[39:39] system learns from what you tap and
[39:40] watch, but you've never been able to
[39:41] just tell it what you want. He credits
[39:44] large language models for making it
[39:46] possible since older ranking systems ran
[39:48] on data that no human could read, while
[39:50] LLMs can now sort content into plain
[39:53] language topics. For now, it is limited
[39:55] to topics, but Instagram says that
[39:58] controls for specific people, moods, and
[40:02] content types are coming.
[40:09] I mean, I think it's a good thing.
[40:11] >> It It seems like what it always like
[40:14] should have had the ability to do. I I
[40:16] recognize the I guess the the challenges
[40:20] around like metadata and tagging and and
[40:24] like how manual something like that
[40:26] could be. Um I can see how LLMs could
[40:31] help with that. But also, I mean, we've
[40:33] spent 90% of the show so far talking
[40:36] about how fundamentally flawed they are
[40:39] and how often they make mistakes.
[40:41] Uh,
[40:43] this is I
[40:47] this will be abusable. I don't know if
[40:49] there will be a reward in abusing it,
[40:52] but this feels like the it is going to
[40:55] be susceptible to like Spider-Man and
[40:57] Elsa style issues. For people that don't
[41:00] understand that reference, kids YouTube
[41:03] years ago was taken over by some really
[41:06] weird content um of people like
[41:09] cosplaying as Spider-Man and Elsa and
[41:11] other characters and doing some weird
[41:14] stuff that you wouldn't really expect to
[41:16] be on the kids side of YouTube. Um but
[41:18] it it got there because of
[41:22] the framing of the video and I could
[41:25] imagine that something similar could
[41:27] happen in this type of system. Um, I
[41:30] think the reward is
[41:33] >> I mean the reward of Spider-Man and Elsa
[41:35] >> was just
[41:35] >> was always
[41:36] >> AdSense.
[41:37] >> Well, I guess yeah, AdSense. So maybe
[41:39] yeah, more reward than for just getting
[41:41] views on Instagram, but there's also the
[41:43] there's also the reward of just getting
[41:46] clout and building up followers on
[41:48] Instagram. Social media accounts are
[41:50] worth money. Um,
[41:52] >> yeah,
[41:54] >> and especially, you know, with
[41:56] automation tools, the the ability to to
[41:59] create accounts to abuse these systems
[42:02] to grow to some threshold to then be
[42:05] farmed off to, you know, scammers who
[42:08] need plausible looking Instagram
[42:09] accounts or, you know, whatever the
[42:10] model is. I I often find myself very
[42:13] surprised by the things that scammers
[42:16] will come up with. Like there was that
[42:18] recent one where there was that like
[42:20] warehouse of consoles that was farming
[42:22] like FIFA card packs or something. I
[42:24] can't remember the exact details of it.
[42:25] I'm not a I'm not a FIFA guy. Dan, do
[42:27] you remember?
[42:29] >> Oh, okay. I just assumed you were into
[42:31] football being an English person and all
[42:32] of that.
[42:33] >> Oh, cricket like a normal person.
[42:35] >> Got it. That makes sense.
[42:36] >> I don't go outside.
[42:40] >> Uh yeah, I remember that topic. I don't
[42:42] remember the specific details. I think
[42:44] it was like farming throwaway matches
[42:46] basically in order to get uh pack
[42:48] openings, like free generating pack
[42:50] openings. I don't know why that benefits
[42:52] them or anything, but yeah, it's it's
[42:55] interesting. I I can't necessarily
[42:56] imagine how you would abuse this. It
[42:58] just feels to me uh how an LLM would
[43:02] view this content
[43:05] is probably abusable in some way. I
[43:08] don't know what to what end, but it
[43:10] feels like it probably would be. Um, I
[43:12] also don't know like what kind of
[43:14] they're even saying you can sort by
[43:15] moods, which is interesting. So, like is
[43:17] it is it doing the Lionus maneuver and
[43:19] is it reading the comments? Um, so could
[43:23] could you manipulate this setup by
[43:28] convincing your audience somehow to all
[43:32] comment in a particular way? Um, I don't
[43:35] know. This almost feels like something
[43:36] like Spiffing Brit should dive into,
[43:38] although he seems to just do YouTube
[43:40] stuff. um
[43:41] >> beast.
[43:43] >> Yeah, but it'd be interesting. I mean, I
[43:47] I like the idea
[43:49] of being able to more manually control
[43:51] our feeds. I I I also respect Lionus'
[43:53] dedication to keeping his Facebook
[43:55] Marketplace weird as heck. Um but I I if
[44:00] if I could just tell Facebook
[44:02] Marketplace that I want it to show me
[44:04] weird things, that would also be nice.
[44:06] Um
[44:07] >> yeah, like one thing
[44:09] >> I think this is good. One thing I've
[44:10] wanted from YouTube for a long time is
[44:14] the ability to tell them what my video
[44:17] is about a little bit because they can
[44:19] mostly figure that out on their own. But
[44:21] what I really want is to tell them who
[44:23] it's for. Because remember remember way
[44:27] back when you made uh here I'm actually
[44:29] just going to pull up the video. Uh come
[44:32] on
[44:35] here. This one
[44:36] >> you might be able to imagine. Oh, for
[44:38] crying out loud.
[44:42] This one right here.
[44:44] >> Installing a video card. How to
[44:46] >> basics on Lionus Tech Tips 10 years ago.
[44:50] >> This video got absolutely dogpiled out
[44:55] of the gate.
[44:55] >> Hated it. People were like actually
[44:57] pissed when they launched.
[44:59] >> So angry. and and and it there was
[45:02] momentum around it because that was back
[45:04] when the like dislike ratio was visible
[45:07] publicly. So people
[45:08] >> it was very negative
[45:09] >> saw that people hated it and then they
[45:12] piled onto it even more and it it
[45:14] created this like like death spiral
[45:16] spiral of hatred and negativity around
[45:18] this video and
[45:21] if you actually just you know watch the
[45:24] video you don't say anything wrong.
[45:28] Hold on. Is it?
[45:30] >> Yeah, it was it was people hated it just
[45:32] because it was beneath most of the
[45:35] average viewers of LT
[45:37] >> at the time.
[45:37] >> Um, yeah, it it wasn't that the
[45:40] information was incorrect. It's just
[45:41] that they were like,
[45:43] >> "It's a graphics card, bruh."
[45:45] >> Yeah, you just put it in.
[45:47] >> But I I'm pretty sure we included things
[45:48] like dduing your drivers and like stuff
[45:50] like that. Like I think it was I think
[45:52] it was like decently
[45:53] >> No, you did a good job
[45:54] >> presented at the time.
[45:55] >> You did a very good job. It was it was
[45:57] lambasted because it hit the wrong
[45:59] audience like you're kind of saying.
[46:01] >> Exactly. So, what I've been asking
[46:03] YouTube for is the ability because
[46:05] because what's happened is we've kind of
[46:08] our channel has kind of Michael Bay
[46:10] itself because we can't just make a
[46:13] video on installing a CPU anymore
[46:17] because what's going to happen is it you
[46:21] bumped your mic. Careful. Uh what's
[46:23] going to happen is it's going to jump
[46:26] into the feed of everyone from someone
[46:29] who subscribed to LTT yesterday to
[46:32] someone who subscribed to LTT 17 years
[46:35] ago. And these folks, your your regular
[46:39] viewers are going to feel like like
[46:42] patronized and they're going to feel
[46:44] like this is a complete and utter waste
[46:46] of time. And what tends to happen is
[46:49] algorithmically YouTube will kind of go,
[46:51] well your channel is about, you know, X
[46:54] and the people who are interested in X
[46:57] don't like this. Therefore, this is
[46:59] probably bad content. But what I would
[47:03] actually really like is the ability to
[47:05] tell YouTube, hey, this video is not for
[47:09] my regulars. This is not
[47:11] >> This was a new viewers like only
[47:14] basically
[47:15] >> and I'm okay with that resulting in it
[47:18] potentially doing worse, especially out
[47:19] of the gate, but I don't want it to
[47:22] massively negatively affect the
[47:24] long-term potential of the video by it
[47:26] nuking a bunch of potential views on
[47:28] people who would never click on it. Cuz
[47:30] like if you know if someone's in flow
[47:32] plane chat or or they've just been you
[47:34] know I I'll meet people at events all
[47:36] the time that are like I've been
[47:37] watching since the NCIX days. like,
[47:39] "Bro, you know how to install a GPU."
[47:41] >> Exactly.
[47:42] >> Like, it's you don't you don't need
[47:44] this.
[47:45] >> Um, the the vast majority of people who
[47:47] have watched more than, you know, one to
[47:51] three LT videos or or realistically any
[47:54] other uh computer hardware tech videos
[47:56] on YouTube, um, isn't going to need a
[47:58] video on how to install a GPU. But that
[48:01] doesn't mean that a bunch of people
[48:04] don't need that. Um,
[48:05] >> so here's another great example. the
[48:07] first time you've ever built a computer.
[48:09] If you just spent a bunch of money on
[48:11] hardware and you want to build the
[48:12] computer for the first time and you
[48:13] don't have an inerson mentor, you want
[48:15] to make sure that you don't ruin this
[48:16] incredibly expensive component as you
[48:18] put it into your computer, you might
[48:20] want to watch a little video on how to
[48:21] install it.
[48:21] >> Yeah. And so we we wanted this was that
[48:24] was supposed to be a series. That's why
[48:26] the title is appended with this how-to
[48:29] basics. I had in my mind that I wanted
[48:31] to create three sort of series of like
[48:34] to really get back down to tech tips and
[48:37] do like how to basics which was kind of
[48:40] a like a play on how to basic which was
[48:43] like kind of a viral thing at the time
[48:44] and then I wanted how to intermediate
[48:47] and like how to advanced. So we would do
[48:50] these tutorials on everything from
[48:51] installing a graphics card to like you
[48:55] know reconfiguring uh ports on your
[48:58] router to uh you know setting up a
[49:02] setting up a a NAS with you know
[49:05] multiple tiers of high speed and low
[49:06] speed storage and we'd kind of we'd kind
[49:08] of segment them out and just create
[49:10] these like very evergreen guides. you
[49:12] know,
[49:14] >> sorry,
[49:14] >> the reception to it was so negative that
[49:17] we did a couple of them and we kind of
[49:18] went, well, holy crap, we can't we can't
[49:21] do this. And then ultimately, years
[49:24] later, that like dislike ratio actually
[49:26] totally turned around and the video now
[49:28] has, you know, 2.3 million views. But
[49:31] that's not nearly as many as it probably
[49:32] would have if it hadn't had its growth
[49:35] stunted right at the beginning there.
[49:39] >> Yeah. It's you, you know what's kind of
[49:41] interesting is I've heard that the
[49:43] youngans these days um to to go back to
[49:47] a previous topic actually also hate
[49:49] Google search now. Uh like Google search
[49:52] just sucks these days. The the results
[49:54] are terrible and the AI search is worse
[49:56] than the auto stuff that they had in the
[49:58] past. Like it's just Google search is
[50:00] just going completely down the hole. Um,
[50:03] but a lot of the youngans are using like
[50:05] Tik Tok for search,
[50:07] which
[50:09] >> hurt my brain a lot when I first heard
[50:11] it. Um, but apparently it's because
[50:16] they're they're using it in the way that
[50:18] a lot of us probably use YouTube, which
[50:22] is where we we look it up like, you
[50:24] know, how do I fix my washing machine
[50:28] >> or whatever.
[50:28] >> Yeah.
[50:29] >> Um, here's a weird question. Someone in
[50:32] someone in chat said that you can
[50:34] Instagram has like a trials feature
[50:36] where it won't send it to your
[50:38] subscribers or something. I I lost the
[50:41] message. Uh Andrew Candy, Instagram can
[50:43] kind of do this with trial reels. It
[50:45] doesn't send them to your followers. I
[50:47] don't know exactly how that works, but
[50:48] maybe the how-to basics series, maybe
[50:52] with a different name, not sure, could
[50:53] return, but in shorts format. I could
[50:56] see potential there.
[50:57] >> This card, but the the shorts version,
[51:00] how to install RAM, but the shorts
[51:03] version.
[51:04] >> It can also even work in long form
[51:05] though. Uh here, Dan, I'm just going to
[51:07] see if my laptop is working. Oh, yeah.
[51:09] There we go. So, I'm on my uh I'm on my
[51:11] daily driver here. So, I'm logged in.
[51:13] So, here's an example of one that kind
[51:16] of shows
[51:17] how bad this can be. Uh so we we we
[51:23] uploaded uh like just building a PC you
[51:26] know just a build guide
[51:28] >> for the first time in like ages. So
[51:31] first day performance of this video was
[51:34] 458,000 literally half of the upper end
[51:38] of our typical range like absolute
[51:41] bombed out of the gate. And then
[51:46] eventually
[51:48] the algorithm figured out, oh, okay,
[51:50] yeah. Oh, this is a highquality video
[51:52] that just the subscribers to this
[51:54] channel were not interested in.
[51:58] Now it's more than double the high-end
[52:01] of your typical range. And just, you
[52:04] know, daily as the English- speakaking
[52:06] world awakens and slepts and awakens
[52:11] again, it just gets views. Um,
[52:16] but it would be really, but it doesn't
[52:17] always work out that way. So, right, the
[52:19] other one that I was going to pull up
[52:21] was Riley did a $1,000 gaming PC. So,
[52:26] here we'll go back to Linus laptop. This
[52:28] one also bombed out of the gate in spite
[52:30] of the very likable Super Riley and then
[52:36] never found that audience. And you know
[52:39] what? Maybe we just, you know, maybe
[52:41] just didn't, you know, work. Maybe a
[52:43] $1,000 guide was not as good as a $2,000
[52:46] guide. Whatever. I don't know the exact
[52:49] answer to why one of them went and one
[52:50] of them didn't. But I've I've been
[52:52] telling Google for years. I would just
[52:54] love to be able to just tell you. I can
[52:58] tell you. I know who it's for. And as a
[53:01] as a variety channel, it's something
[53:03] that we've struggled with our entire
[53:05] existence. And I think we struggle with
[53:09] more and more now where one video on our
[53:12] channel one day is touring a data center
[53:16] and then the next day is like tearing
[53:19] down the walls of the tech house so that
[53:21] we can run conduit and and low voltage
[53:24] wiring and then the next day is a review
[53:27] of a steam controller. like we're all
[53:29] over the place. And I have I have
[53:32] noticed that algorithmically they just
[53:35] don't know what to do with us compared
[53:38] to a channel that is more more
[53:40] predictable. Uh like even within the
[53:43] tech niche, there's other channels that
[53:44] are more predictable in terms of their
[53:47] their tone from one video to the next.
[53:49] And I'd love to be able to just say,
[53:51] "Hey, this is for people that are super
[53:54] into biohacking or, you know, this is
[53:59] for people who are really into, you
[54:00] know, enterprise hardware, you know,
[54:02] whatever right?"
[54:05] >> But I can't.
[54:06] >> Yeah.
[54:07] >> There's breaking anthropic news, by the
[54:08] way, that we should probably acknowledge
[54:10] because chat is gloating.
[54:14] >> I'm working on that right now. Um
[54:18] >> basically they issued a statement um
[54:22] >> the US government shut them down.
[54:25] >> Statement on the US government directive
[54:27] to suspend access to Fable 5 and Mythos
[54:30] 5.
[54:32] Um
[54:34] yeah, citing national security
[54:37] authorities, the US government has
[54:38] issued an export control directive to
[54:41] suspend all access to Fable 5 and Mythos
[54:43] 5 by any foreign national, even those
[54:46] inside the United States, including
[54:50] anthropic employees.
[54:54] Uh
[54:58] so
[54:59] >> this feels like
[55:01] Yeah. To ensure compliance, they have
[55:03] just turned Fable 5 and Mythos 5 off
[55:08] for everyone.
[55:10] >> Uh, this Dan Dan pointed this out in
[55:13] full chat. Uh, but this is probably the
[55:16] best marketing anyone on the entire
[55:17] planet could have asked for. But I
[55:19] actually think this might be hilariously
[55:24] the whole marketing angle might not be
[55:27] the goal that the US government wanted
[55:29] because I suspect they are still in a
[55:31] bit of a tiff over when Anthropic stood
[55:34] up to them a little while ago. Um, and
[55:38] with with the IPOs coming soon, I
[55:42] suspect this might have been uh an
[55:45] attempt to hold them back actually. But
[55:47] I think just like last time, this is
[55:50] ultimately going to be just like an
[55:52] insane marketing move. Uh, promoting the
[55:56] absolute heck out of anthropic. Um,
[56:01] the the the including internal employees
[56:03] thing is is wild. We we had um Conrad,
[56:07] the the guy who makes LT store. Yeah. In
[56:10] regards to development, um was in chat
[56:12] earlier saying he used Fable 5
[56:15] personally um and said I I don't I don't
[56:18] want to misquote. I don't remember
[56:19] exactly what he said, but he said very
[56:20] positive things about it. Um
[56:24] very interesting. Honestly, um
[56:28] the Frontier models that are coming out
[56:30] right now are pretty nuts. Um
[56:35] yeah, Mari squared in fulloint chat said
[56:38] not me currently using Fable 5 to
[56:40] rewrite a whole platform. Um
[56:44] yeah wild
[56:46] wild
[56:48] in in before uh US citizens start
[56:52] reselling uh their their access to Fable
[56:55] 5 to foreign nationals. Um,
[56:57] >> yeah.
[57:00] >> If only there was if only there was some
[57:02] technology that could make it so that
[57:05] you could appear as though you were in a
[57:07] different region than you were. If only
[57:09] there was some way to pay for things
[57:12] with a, you know, US bankisssued payment
[57:16] method. If like this, I don't know how
[57:18] they could possibly hope to to enforce
[57:21] this once they actually do reenable
[57:24] Fable 5. Yeah, that's what I was going
[57:26] to say is like I I don't think it's
[57:27] going to be as simple as as a VPN, which
[57:29] is why they actually shut it down.
[57:31] >> No, but you will but you'll need but it
[57:33] won't someone will overcome it. This
[57:35] ain't going to be impossible to overcome
[57:37] no matter what.
[57:39] >> At the very least, I suspect it's going
[57:40] to be people individually reselling API
[57:43] key usage um somehow. But it's yeah,
[57:50] wild wild. Verified US ID requirements
[57:53] incoming. Maybe. Yeah.
[57:55] >> Okay. I'm not going to say who said
[57:57] this, but somebody said, "I got a notice
[58:01] on a litigation hold on all
[58:02] communications about anthropic. M DoD
[58:05] employee."
[58:07] I mean, I think you mean DOW.
[58:10] Um, but yeah, for now DoD.
[58:13] >> I've I've heard I've heard the
[58:15] Department of War thing
[58:17] isn't official. They just call it that.
[58:20] No, but it's it like went looking like
[58:23] probably official today or yesterday or
[58:26] something. Yeah,
[58:27] >> got it. Got it. Got it. Got it. Um I I
[58:31] don't I don't know what's going on with
[58:32] that, but thought that was kind of
[58:35] funny. Has to be approved by Congress.
[58:37] Okay. Yeah.
[58:38] >> Yeah.
[58:39] >> Um yeah, wild situation. Absolutely wild
[58:41] situation. I I think like um I don't
[58:44] know if we've talked about it on W show
[58:45] or not um but but PewDiePie's video on
[58:49] Odysius.
[58:50] >> Yeah.
[58:51] >> Um
[58:52] it's it's you know we've talked about
[58:55] PewDiePie a few times over the years on
[58:57] W show. Uh this is not one of the ways
[58:59] that I expected to be talking about
[59:00] PewDiePie, but he has been
[59:04] doing incredible things when it comes to
[59:06] locally hosted AI and Agentic stuff. and
[59:09] his video on Odysius uh his his project
[59:14] uh which he has open sourced um has 3
[59:17] million views now. And you know we
[59:19] talked about this last show but u you
[59:22] know local models are not are not
[59:25] necessarily you know brushing shoulders
[59:27] with these frontier models. Um,
[59:30] >> no, but they're right now,
[59:33] >> but they are very useful. And I really
[59:36] think the in the same way that with with
[59:39] the whole operating system debate right
[59:41] now, my solution is to very happily just
[59:43] dual boot on both of my systems. I'm
[59:45] dual booting on my laptop, I'm dual
[59:47] booting on my desktop, and I'm very
[59:49] happy. Um, I think the like prime
[59:52] solution for me at least is going to be
[59:54] using both. um saving your API costs on
[59:59] something big and heavy and powerful
[60:02] like uh Fable 5 or or any other like
[60:05] crazy frontier model because you know in
[60:07] two weeks it'll probably be something
[60:09] else that's in the lead. Um and then for
[60:12] most of your
[60:15] honestly probably most of your work but
[60:17] it really depends on who you are and
[60:18] what you're doing. Um, you can use local
[60:20] models in order to
[60:22] >> just like your your just dumb [ __ ]
[60:24] like what's the
[60:26] >> what's the weather going to be tomorrow
[60:28] just to to interact with, you know, your
[60:31] your DIY Google Home that doesn't
[60:34] actually communicate with Google's
[60:35] servers. Uh, like that that cute little
[60:38] robot thing that Nick Harris from the
[60:40] lab had in his AMD ultimate tech
[60:42] upgrade. Like paying for tokens to run
[60:45] something like that to turn your lights
[60:46] on and off is cra would be crazy.
[60:49] And
[60:50] >> yeah, there's no need to
[60:51] >> and you Yeah, you just don't need to
[60:53] because a local model can handle the
[60:55] voice to text can can can handle the
[60:58] natural language interaction just fine
[61:00] today. Uh
[61:02] >> there was that guy in chat who said they
[61:04] were they were using uh Fable to rewrite
[61:07] an entire system and it's like okay.
[61:09] >> Yeah. Um
[61:10] >> you're not doing that on your RTX 3070.
[61:14] >> Uh yeah, okay, sure. um you you you can
[61:16] you can do quite a bit with local models
[61:18] and and it might do pretty well. But
[61:20] >> yeah, Atomica says Quinn 3.6 is
[61:23] phenomenal
[61:24] >> and that happens to be uh that was
[61:26] actually one of the first things that
[61:29] Nick Harris asked when we were meeting
[61:31] with Nvidia at Computex was okay neat
[61:36] impressive demo you what are you guys
[61:39] using and Quen 3.6 came up like
[61:42] immediately. Uh it seems like it's just
[61:45] basically beloved right now for anyone
[61:48] who wants to run local AI.
[61:51] >> Yeah, it's uh there there's a lot of
[61:53] different things. If you're if you're
[61:55] scared of running stuff, again, I would
[61:56] I would highly recommend watching the
[61:58] PewDiePie video, but he talks about a a
[62:00] part of the Odysius system called
[62:01] cookbook, uh where it helps inform you
[62:05] based on your system hardware, uh what
[62:07] what you can run. So, if you're like,
[62:09] "What the heck is a what the heck is a
[62:11] Quen?" um it can it can kind of handle
[62:14] those portions for you or at least
[62:17] majorly simplify it. But yeah, the the
[62:19] local hosting stuff is was incredibly
[62:21] interesting and is now
[62:24] taken a massive leap forward in how
[62:26] interesting it is and the Frontier
[62:28] models keep being mind-bogglingly
[62:31] insane. Um and it's it's quite quite the
[62:35] time for uh terrible things to happen.
[62:39] Um,
[62:41] yeah. Anyways, topic move on.
[62:44] >> Oh, uh, no, it's time for us to do our
[62:46] creator warehouse announcements. This
[62:49] week, we are launching a pretty fun one.
[62:53] I am legitimately super excited about
[62:56] this one. Okay, hold on. Hold on. It's
[62:58] time to hopefully not have a classic
[63:01] Wancho nip slip here. Okay.
[63:05] Oh, shoot. I still have my mic on. Uh,
[63:07] okay. First up is the map of Siberia
[63:12] t-shirt. It is a fully handustrated
[63:16] fantasy map where someone replaced all
[63:20] the kingdoms with things like
[63:23] >> the great firewall
[63:25] is of when and the cryptocurrency.
[63:30] Get it? Cryptocurrency.
[63:34] >> That's pretty good.
[63:34] >> Hold on. Hold on. I'm going to go to the
[63:36] uh to the Lionus cam. Uh this is got to
[63:39] be Oh man, you guys can't Okay, you know
[63:42] what? I'm just going to have to I'm just
[63:43] going to have to pull it up on the site.
[63:45] Um
[63:47] >> uh yeah, I I'll bring it up on the site.
[63:48] I'll bring it up on the site. Uh I I
[63:51] have to give Lisa from the CW team like
[63:56] the vast majority of the credit on this
[63:59] one. this. She spearheaded this and just
[64:03] oh, I don't know when it happened, but
[64:09] she just like figured out our vibe as a
[64:14] company, our audience at some point.
[64:18] >> Pirate Bay
[64:19] >> and it culminated or maybe she's still
[64:22] cooking. Maybe she maybe she hasn't
[64:24] reached the peak yet, but oh my god, the
[64:28] leg access point. You have got to be
[64:31] kidding me.
[64:32] >> That is the funniest [ __ ]
[64:36] >> Control + C is really good. I like that
[64:38] one.
[64:39] >> Yeah,
[64:40] the poisoned cash
[64:42] >> port 80
[64:43] >> network bridge.
[64:47] Of course, this had to be there. This is
[64:50] This is This is quite informed actually.
[64:53] >> Right.
[64:55] >> The bits.
[64:56] >> That's cool.
[64:57] >> The bites.
[64:59] I asked her. I was like, "Are do the
[65:02] bites have exactly eight times the land
[65:04] mass of the bits?" And she was like,
[65:05] "Don't overthink it." I'm like, "Okay,
[65:07] all right. Fair fair enough, Lisa. All
[65:09] right. All right. I guess
[65:10] >> Hold on a second. I got
[65:11] >> the data
[65:12] >> 27032."
[65:16] >> Oh, interesting.
[65:17] >> Interesting. Don't don't don't worry
[65:18] about it. Don't worry about it. Um,
[65:21] absolutely
[65:23] absolutely incredible design. Super
[65:26] cool.
[65:27] >> Are these up right now?
[65:29] >> Yeah.
[65:30] >> Okay. I didn't see a link in the doc.
[65:32] >> Map of Siberia.
[65:33] >> Oh, I see. There it is.
[65:34] >> T-shirt available now in regular and
[65:37] tall. So, make sure you guys are getting
[65:39] in there. Oh, we don't have a lot of
[65:40] tall units in the uh the XL's, double
[65:43] XXLs, and triple XL's, at least on the
[65:45] global site. And great photo shoot as
[65:48] always. This has to be Tynen.
[65:51] There's no way this isn't Tynen. It's
[65:53] Tynen. No, it's Dan.
[65:55] >> That's me.
[65:56] >> Did Did you borrow T? Okay, this one's
[65:58] Titan.
[65:59] >> The ones outside are Tynen. Yeah.
[66:01] >> Okay.
[66:03] Amazing. I know he's into that stuff. I
[66:05] didn't know you were. Did you borrow his
[66:06] stuff?
[66:07] >> Uh, it was actually somebody else's
[66:08] stuff.
[66:09] >> Oh, really?
[66:09] >> There was some other guy who lent us a
[66:12] set of armor.
[66:13] >> Oh, that's super cool. Um,
[66:15] >> hold on one sec. Am I Am I getting this
[66:17] correctly? The olive one has the large
[66:20] map on the front,
[66:21] >> but the black one has the small logo on
[66:23] the front, the large map on the back.
[66:25] >> Yeah. So, we couldn't decide uh we
[66:28] couldn't decide whether to do olive or
[66:30] black or the here, if I switch over to
[66:33] the black, you'll see this one has the
[66:35] map on the back and then it has the uh
[66:39] just the cloud kingdom on the front. So,
[66:42] we kind of tried both and we'll kind of
[66:46] see what people are more into.
[66:49] I'm not going to spoil anything by
[66:50] saying who thought which way, but uh
[66:53] they're they're both really cool designs
[66:56] and
[66:56] >> I'm so interested how the sales of this
[66:59] go. I I'm so interested because I've
[67:01] always personally like I I think this
[67:04] shirt is wicked. I'm going to try, if
[67:06] there isn't one sitting there for me
[67:08] already, I'm going to try to get the the
[67:09] black one in XL tall. But, um, the
[67:12] reason why I'm going to get the black
[67:14] one is because I always love the, um,
[67:19] the small logo over the heart with Is
[67:24] that Is that my size?
[67:26] >> Sorry, it's a medium.
[67:28] >> Darn it. Uh, I always really like the
[67:30] small logo over the heart and then the
[67:31] big thing on the back of the shirt.
[67:33] That's always my favorite shirt design.
[67:36] But I'm wondering
[67:40] is that is that just a me thing or is
[67:42] that
[67:42] >> No, it's a lot of people thing. But we
[67:44] also felt that the way that this design
[67:47] doesn't pop as like Okay, here.
[67:50] >> No, I agree. That is a good point.
[67:52] >> Like this is a lot to wear on the front.
[67:56] This is actually not as much.
[67:59] >> Yeah. No, I I agree. It's very
[68:01] interesting. Yeah.
[68:02] >> And it's also not like a huge logo. Um
[68:07] if if it was a massive like your entire
[68:10] torso Linus Tech Tips logo.
[68:12] >> Yeah,
[68:13] >> that would it probably kind of suck.
[68:15] >> But instead, it's like a cool design
[68:17] full of fun little Easter eggs and stuff
[68:19] like it's
[68:20] >> and there's the you know there's the the
[68:22] shirt that must not be named uh that
[68:24] doesn't exist anymore. Um and that had a
[68:27] big logo on the front. Sleep is fories.
[68:30] >> No, it was released very recently and
[68:32] then taken away.
[68:34] >> Oh. Oh, that one.
[68:36] >> Yeah, that had a big logo on the front
[68:38] and that looked really sick as well. So,
[68:40] like it's obviously not a hard and fast
[68:42] rule.
[68:42] >> So, I am interested because just like
[68:44] you said on the olive one,
[68:46] >> it it does it kind of works. Um,
[68:49] >> and um,
[68:51] >> uh, well, one that I do have an XL tall
[68:53] here for you is this one was our best.
[68:56] >> See, that also works.
[68:58] >> This was our bestselling t-shirt in a
[69:00] long time. Um,
[69:02] >> so it's really not a hard and fast rule.
[69:04] It's just like a generalization that I
[69:06] prefer that setup.
[69:07] >> But like I wouldn't that pirate one
[69:10] >> if you just had a little parrot. It's
[69:12] not showing up on camera. If you just
[69:13] had a little parrot here and then the
[69:15] back said that,
[69:17] >> I don't think that would be better. Um,
[69:19] >> no, I don't think that one would quite
[69:20] work. So, this is like
[69:22] >> kind of depends.
[69:23] >> We've been kind of trying to find our
[69:25] stride um on like the t-shirt designs
[69:29] over the last little while. Um, like I
[69:33] think we had um I don't want to get too
[69:35] deep into like sort of the the internal
[69:40] stuff, but you guys might have noticed
[69:42] that whether it's sourcing the blanks or
[69:45] whether it's having designs ready or
[69:47] whether it's having both of those things
[69:49] at the same time, um, we we haven't
[69:52] really focused on t-shirts for like
[69:56] a couple of years at this point. Whereas
[69:59] it used to be that like every other
[70:00] freaking Wand show we were announcing
[70:02] another t-shirt and t-shirts were like a
[70:03] big part of LTT store and then for a
[70:06] long time they basically haven't been.
[70:08] But um I think over this last couple
[70:12] like I I obviously haven't seen any
[70:14] sales figures for this one yet but I
[70:16] love it. Uh the design team freaking
[70:18] loves it. Like we love it so we're
[70:20] really confident in it. But as you can
[70:22] see, we're still kind of mixing it up in
[70:24] terms of, okay, do we do little thing on
[70:26] the front, big thing on the back, big
[70:28] thing on the front? Like, what's the,
[70:30] you know, what's our what's our current
[70:31] meta going to be? So, we're going to
[70:33] we're going to let you guys
[70:34] >> ultimately decide that.
[70:36] >> Um,
[70:36] >> it's very interesting. It's I I'm I'm
[70:38] specifically interested in the sales of
[70:40] this one with my little I like the small
[70:42] logo thing. Um, because I think this one
[70:46] is a strong example of the big image on
[70:48] the front being cool. Yeah. Um,
[70:52] >> so it's it's it's like a a good test.
[70:54] Um, both are cool, I think.
[70:57] >> And but wait, there's more.
[71:01] >> Where is it? There's also Oh, our new
[71:03] arrivals could probably stand to be
[71:05] cleaned out a little bit. Whoops.
[71:07] >> A desk pad. Is that what you're talking
[71:08] about?
[71:08] >> Yeah. So, there's also a desk pad. How
[71:11] fun is that? Available in three
[71:15] different sizes. It's 900 mil by 400
[71:18] mil, 300 mil, or 600 mil depending on
[71:21] your preference. All the same price as
[71:23] you've come to expect from LTT store.
[71:26] Hello from LTT store. It also looks
[71:30] extremely striking. Great quality
[71:32] printing. Got this here. Map of Siberia
[71:36] desk pad. Oh, it's Luc is there. Oh,
[71:40] yeah. There you go. Freaking awesome.
[71:42] And of course, nice high quality
[71:44] deskpad. natural rubber stitched etches,
[71:46] etc., etc. Stitch etches. Stitched
[71:48] edges. What you would expect from us,
[71:52] man. They've been having way too much
[71:54] fun at these photo shoots lately.
[71:58] >> And that's a good thing. One of what?
[72:05] >> Oh, really? Unless it's not on there. I
[72:08] didn't see that one. Sorry, Dan. I
[72:11] didn't see it. Anyway, those are our new
[72:13] launches for this week and you can check
[72:16] them out at lmg.gg/cyiberia,
[72:19] which is spelled like cyber
[72:22] and then I A.
[72:25] Um, and it's a perfect time to pick one
[72:28] up because we're live, which means that
[72:30] you can send a comm.
[72:33] We might need to change the name again.
[72:35] Well, whatever. For now, they're called
[72:36] checkout messages coms. And the way that
[72:39] they work is all you got to do is fire
[72:40] up LTT store, add something that catches
[72:43] your eye to your cart, boop boop, view
[72:46] your cart, and in the checkout, you can
[72:48] type up a message. It'll go to producer
[72:50] Dan. There he is. Who will respond to it
[72:54] or just like pop it up on the stream if
[72:56] it's just like a shout out for someone
[72:57] or he will curate it for me and Luke to
[73:01] respond to. Uh Dan, do you want to hit
[73:04] us with a couple of coms? Sure thing.
[73:07] Lionus, your 40th level is coming up. Do
[73:10] you plan on celebrating it in a special
[73:11] way? How about writing a biography or
[73:13] even an autobiography? I'd read a
[73:16] success story. Also, more non-black
[73:19] t-shirts, please. Um, I'm flattered, but
[73:23] I
[73:25] am
[73:27] not planning to write an autobiography.
[73:31] I I started making notes like a couple
[73:33] times for something like that and then I
[73:36] just have not had the time to do it
[73:39] slashh
[73:44] I don't know I don't maybe maybe I've
[73:45] just been maybe I've just not thought
[73:47] about it the right way cuz I was I was
[73:49] really surprised when I found out that
[73:50] someone as private as Linus Torvalds had
[73:53] written an autobiography and when I
[73:56] asked him about it um he basically was
[73:58] just like well try anything once
[73:59] including
[74:03] write an autobiography apparently cuz
[74:06] he's a really modest guy and that didn't
[74:08] really he's a really modest really
[74:10] private guy and that didn't really align
[74:12] with write an autobiography
[74:15] but the writer actually talks about this
[74:18] in the autobiography um or I guess it's
[74:21] not really an autobiography but it's
[74:22] like it's got a lot of autobiographical
[74:26] elements but also a lot of observations
[74:29] from the author. It's kind of like a a
[74:30] mish mash of both. Um, and the writer
[74:33] actually talks about the process by
[74:34] which he convinced Torvalds to do it.
[74:37] Um, so maybe I don't know, maybe I just
[74:39] haven't heard the right pitch from a
[74:40] ghost writer. I also just don't think
[74:41] I'd be comfortable with a ghostriter
[74:43] writing it. I feel like I should write
[74:44] my own thing. I literally people ask me
[74:46] like, "What do you do?" And more than
[74:48] anything else, I think more than a host,
[74:50] um, I I think my greatest contribution
[74:53] to LTT is is writing. Uh, I don't think
[74:56] people realize how much of what is done
[75:01] at Lionus Media Group has been touched
[75:03] at some point by my pen. Um, not all of
[75:07] it.
[75:08] >> It's all of it.
[75:09] >> It's not all of it.
[75:10] >> Every single mistake.
[75:11] >> Thank you.
[75:12] >> It's all Lionus.
[75:13] >> Thank you for that.
[75:14] >> Even if it's just a Labs article that
[75:16] doesn't have him credited, if there's a
[75:17] mistake in it, it was Lionus. If there's
[75:19] not a mistake in, it was someone else.
[75:24] All right.
[75:26] >> Anyways, uh
[75:28] >> did I answer the question? I don't
[75:29] remember and I I'm having a hard time
[75:31] logging into the dashboard.
[75:32] >> No, I think that's probably good enough.
[75:34] >> All right, cool. Good chat.
[75:35] >> Hey, DL. What's the most recent book
[75:37] you've read or are currently reading?
[75:40] >> Oh,
[75:44] um what did I pick up recently?
[75:47] >> Trying to remember. I read Breakfast
[75:48] with my a little while ago, but that was
[75:50] a while ago. I think I read something
[75:52] more recently than that.
[75:56] >> I'm second here.
[75:58] >> I've I've reverted to my teenage state
[76:01] and I've been uh I've been reading old
[76:03] Foxtrot collections during my midnight
[76:06] snack breakfast cereal lately. So,
[76:08] that's been that's been fun. I love
[76:10] Foxtrot.
[76:11] >> Fun.
[76:12] >> I actually had no idea he still
[76:14] publishes Sunday strips online. Shout
[76:17] out Bill.
[76:21] Um, man, there's no way to go sort by
[76:24] recent. My goodness.
[76:31] I will find it. Give me a second.
[76:33] >> He's looking up Dan in the meantime.
[76:35] Read any good books lately?
[76:37] >> I've still got most of The Prince to
[76:40] finish off.
[76:44] >> Um,
[76:45] >> what?
[76:45] >> I read it back in college and it was
[76:47] like
[76:48] >> it was brutal and it's still brutal.
[76:50] reminds me a lot of the world right now.
[76:53] >> Have you um Have you read it, Linus?
[76:55] >> No, I haven't.
[76:56] >> Yeah, you can have my other copy if you
[76:58] want. It's um it's bad. It's hard. It's
[77:01] hard to read,
[77:02] >> huh? What a what a sales pitch.
[77:05] >> It should be read, I think, by a lot of
[77:07] people but
[77:08] >> Okay, that's a better pitch. You should
[77:09] lead with that.
[77:10] >> The prince is good.
[77:15] Um,
[77:19] >> holy crap. You know what, Lionus? I can
[77:21] give you something fun. Oh, there it is.
[77:24] Um, Tough and Competent by Eugene F.
[77:27] Crayons.
[77:28] >> Tough and Competent.
[77:30] >> Is that Is is that your autobiography?
[77:34] >> No, but thank you. Um,
[77:37] yeah. Uh Jean CR was uh the second uh
[77:43] chief flight director for NASA.
[77:45] >> He looks tough and competent.
[77:47] >> Yeah.
[77:50] Yeah. Yeah. Um but I I haven't actually
[77:54] started it. Uh but I have it and I'm
[77:57] excited.
[77:57] >> That's the first step.
[78:00] >> One small step for Luke's
[78:03] reading list.
[78:06] >> Yeah. Um, but okay, this is okay. I
[78:09] guess is that that question done?
[78:13] >> I'm excited to show you something. And I
[78:14] keep trying to jump to it and then
[78:15] realizing that I'm skipping the topic
[78:17] we're currently on. Sorry.
[78:20] >> Uh, this is not an ad. I don't know if
[78:22] these guys have ever sponsored us or
[78:24] not. So, my bad if they have, I don't
[78:26] know, whatever. But, um, and I can't
[78:28] send you a link to this super easily,
[78:30] but if you just want to Google it,
[78:32] you've probably already seen it. But um
[78:34] 8bit do ultimate three mode controller
[78:37] Xbox. They have a 8bit do Xbox
[78:40] controller that is translucent green.
[78:42] Have you seen this?
[78:44] >> Uh I thought you would like it.
[78:47] >> Three mode. Oh, nice. That looks sick.
[78:50] >> Yeah. Yeah. It's really interesting.
[78:53] There's a bunch of I I don't know if
[78:56] everyone just had the same thought at
[78:57] the same time or what, but there's a
[78:59] bunch of translucent products coming
[79:00] out. You can see in the background
[79:02] there, they have a keyboard. Um, yeah,
[79:04] they have an 85% keyboard and a separate
[79:07] 10 key, which I think is super cool.
[79:09] Those buttons tend to come with their
[79:10] keyboard. Um, so I'm assuming that's the
[79:13] case here as well. And then they also
[79:14] have the mouse. And then they also have
[79:16] the uh the Xbox controller. It's a sick
[79:19] lineup of hardware actually.
[79:21] >> That looks amazing. Officially licensed
[79:23] by Xbox.
[79:24] >> Awesome.
[79:25] >> Supported by Apple, man. 8bit Dough has
[79:27] just been crushed.
[79:29] >> I believe they're Hall effect joysticks
[79:32] or or or something. Um I lost the
[79:36] >> Hall effect joysticks. Yep.
[79:38] >> Yeah.
[79:38] >> And the triggers are Hall effect linear.
[79:41] >> Oh, nice. Okay. The triggers I wasn't
[79:43] sure, but I was trying to find that.
[79:44] Cool. Um and there's a switch on the
[79:46] back to switch which kind of device
[79:47] you're connecting it to and stuff. This
[79:49] looks wicked. Um I I just I I don't
[79:53] know. I'm just so stoked on translucent
[79:55] products coming out. Um, I think it's
[79:58] awesome. I love I'm so happy that it
[80:02] seems like we're exiting an era of
[80:04] everything just being either all black
[80:06] or all white. Um, I'm so happy that some
[80:09] color is entering products again and
[80:11] that translucent and seeing the
[80:13] interesting things inside the product is
[80:16] becoming a thing again. I just think
[80:17] it's so sick. Oh, dude. I um Okay, I had
[80:22] a similar experience today. Uh I was
[80:25] working on the LTT for uh Sony sent over
[80:28] their 115in true RGB TV. So, it's uh
[80:32] it's over in the theater room right now.
[80:34] I I had to leave for WAN show, so I
[80:36] didn't actually get to see it fire up
[80:39] and watch content on it. I have to do
[80:41] that on Monday. I have to wait the whole
[80:42] weekend before I ah Anyway, that's not
[80:44] the point. The point is, I was looking
[80:46] at the remote and it's a black remote
[80:50] and I was like, "Of course it's a
[80:51] black."
[80:53] No wait.
[80:54] >> Translucent.
[80:55] >> It No, no, it doesn't. It has It had
[80:57] little uh like teal like flexcks in the
[81:02] plastic like Do you remember our uh like
[81:04] our RGB shirts how they had those little
[81:07] like like uh
[81:08] >> like those little color tufts in them?
[81:12] >> Yeah.
[81:12] >> Yeah. It was cut. It was all one color,
[81:15] but it was kind of like that. Like it
[81:16] just had little color flexcks in it. And
[81:18] I was like, how fun is that? And how
[81:22] hard was that, you know, to bring back a
[81:25] little bit of
[81:27] color in our lives? I I just I thought
[81:29] it was really cool. I thought it was a
[81:30] nice little touch.
[81:33] >> Yeah. I I think that might actually be
[81:36] um a double benefit because I believe
[81:38] that's often an appearance that comes
[81:40] from recycled plastic.
[81:42] >> Yeah.
[81:43] Um, so it could be a recycled plastic.
[81:46] >> It looks great.
[81:47] >> Oh, I I think it looks fantastic. I
[81:49] think it looks fantastic. Also, I didn't
[81:51] know this was a thing, but just to to
[81:54] be, you know, fair. Um, and I'm doing a
[81:56] terrible job with tracking names today
[81:58] by East One. Um, or Eastone
[82:02] um posted in Flipline Chat. Xbox is also
[82:05] releasing their own translucent green
[82:08] Xbox controller. If you look up the Xbox
[82:11] 25th anniversary, um they have an Xbox
[82:13] wireless controller X25 special edition
[82:16] and it's a translucent translucent green
[82:19] Xbox controller. Um
[82:22] >> I'm a little confused about some of the
[82:24] messaging that I hear coming out of Xbox
[82:26] right now. Like on the one hand, they
[82:28] need to rejuvenate their hardware
[82:30] business, but then on the other hand,
[82:31] they're saying things like how many
[82:33] Xboxes they sell is gated by like
[82:36] production capacity, not by like demand.
[82:39] I just I'm I'm having a really hard time
[82:41] reconciling that. I'm also I am also
[82:44] recognizing that a lot of Xbox fans are
[82:48] rooting for exclusives,
[82:51] but I can't help but feel like if they
[82:54] go full exclusive and they don't come to
[82:57] PC that we are like we're sitting here
[83:00] demanding reverse progress that is
[83:03] actually regression.
[83:05] Um, as someone who has never bought an
[83:08] Xbox and who probably will never bother
[83:11] to game on an Xbox, um, making games not
[83:14] available for me to purchase on PC is
[83:16] something that I'm going to have a hard
[83:17] time rooting for. Uh,
[83:20] >> the Xbox install base being as it is,
[83:23] games being exclusive might just mean
[83:26] their death.
[83:27] Yeah,
[83:29] but I also understand that like there's
[83:32] it's a chicken and egg problem because
[83:34] you're not going to get an install base
[83:36] unless you give people a reason to buy
[83:38] the console and it sure as heck isn't
[83:39] going to be the Xbox first party
[83:42] controller. Like what's it going to be?
[83:44] >> I don't feel like it's going to be Gears
[83:47] 6 either. Um
[83:50] or or whatever.
[83:52] >> I mean, you weren't really like a Gears
[83:53] kid though, right?
[83:55] >> Not really. But
[83:58] I feel I feel like Ooh, is this hot? Is
[84:01] this a hot take?
[84:02] >> Here comes the hot take.
[84:03] >> I feel like Gears always played second
[84:05] fiddle to Halo.
[84:06] >> Well, yeah. I don't think that's a hot
[84:08] take at all, sir.
[84:09] >> And I don't feel like either series is
[84:11] hot right now. So, what uh Halo Infinite
[84:15] the second and Gears 6? Like, I don't
[84:20] know. uh whatever whatever that thing
[84:23] they're doing is with
[84:26] uh
[84:28] the the the first Halo that they're
[84:30] recreating and they're making new
[84:31] missions.
[84:32] >> So
[84:32] >> yeah, um I don't think that's going to
[84:34] sell consoles either.
[84:36] >> I was chatting with Poof about it
[84:38] actually right before the show and he
[84:41] was like super bearish on it and then
[84:44] just saw the latest updates and he was
[84:47] like, "Yeah, I'm psyched now.
[84:51] I haven't actually I haven't seen it,
[84:53] but apparently it looks great
[84:56] >> like visually,
[84:57] >> but then also, didn't they have one of
[84:59] the old Bungie guys come out and pretty
[85:01] much say like, yeah, it looks great, but
[85:04] it doesn't look like Halo Combat
[85:06] Evolved. So,
[85:08] I'm and I'm paraphrasing, I'm basically
[85:10] an AI right now. Disclaimer, what I said
[85:12] might not be attributable to anyone.
[85:14] >> I'm basically an AI right now is a
[85:16] hilarious line. Um,
[85:19] the three new missions is interesting
[85:21] and stuff. I
[85:21] >> don't trust me, bro.
[85:23] >> All I'm saying I'm not saying it doesn't
[85:24] look interesting. All I'm saying is I
[85:26] don't know if it's a console seller. Um,
[85:32] I don't know what Microsoft has right
[85:34] now that would be a console seller.
[85:37] People really like Fors Horizon 6.
[85:44] >> Okay.
[85:45] >> Yeah. They also dropped Japan, which has
[85:47] been a location that people have been
[85:49] looking for in Fors Horizon for a long
[85:50] time. So, they just dropped their like
[85:52] ace up their sleeve
[85:54] >> before it could be an exclusive. Um,
[85:58] I just spent, you know,
[86:01] days driving switchbacks in Taiwan. They
[86:04] could do one in Taiwan, but I I don't
[86:05] think that's going to be as big of a
[86:06] draw as Japan is for car culture,
[86:09] obviously. Um,
[86:14] I don't know. I don't I don't know what
[86:15] game
[86:17] Microsoft has up their sleeve. Oh, yeah,
[86:20] I do. Okay.
[86:24] I mean, if some new Xbox launches and
[86:27] soon after the launch, Elder Scrolls 6
[86:29] comes out and it's exclusive to
[86:32] >> consoles, that'll move some consoles.
[86:35] >> That'll move some cons. Well, assuming
[86:37] it's not a
[86:39] complete dumpster fire, it'll move
[86:41] >> I think it'll still move some consoles,
[86:43] honestly. It's just a question of how
[86:45] many sir.
[86:46] >> Yeah, totally. Yeah, exactly. Um, don't
[86:49] curse us like that. Yeah, sorry. That's
[86:51] a good point, but sorry. Um, happy 8th
[86:54] birthday to the trailer of ES6. I didn't
[86:58] believe that trailer for so long. I
[87:00] thought it was fake. I thought it was a
[87:02] fan edit. Uh, I didn't watch that game
[87:05] show, whatever one it was announced in.
[87:07] So, when I saw it go on YouTube, I I
[87:09] legitimately thought it was just like a
[87:11] fan edit. um and didn't believe it for
[87:14] so long and then people were talking
[87:15] about it like I think years later and I
[87:19] was like wait what that was real
[87:22] um that was wild. I can't believe that I
[87:25] still can't believe they did that. Um
[87:28] yeah I I don't think Oh man I I'm kind
[87:33] of strongly of the opinion that it's
[87:34] been it's been downhill since Morwind.
[87:37] Um, but the peak of Morwind was so high
[87:40] that the slope was still really, really
[87:42] good. Um, but like Fallout 4,
[87:47] even compared to Skyrim,
[87:50] even compared to Oblivion, I think was
[87:52] like a continual
[87:55] down. Fallout 4 was still really fun. It
[87:57] was a great game, but it felt like the
[87:58] worst one out of those. I feel like as
[88:01] someone who came into the series on
[88:03] Oblivion,
[88:04] >> it's tough and then tried to play
[88:06] Mororrowind because Oblivion was so
[88:09] great and I wanted more.
[88:11] >> You can't go back. You You have to look
[88:13] at Morowind in the time capsule of when
[88:15] it was released.
[88:16] >> Sure.
[88:16] >> Um I'm not I'm not suggesting you go
[88:18] back and play it now, but when it was
[88:20] released, Morowind was an insane game.
[88:22] Like there was
[88:24] >> there was nothing like it. Um,
[88:27] it it was wild. Um, and the the the way
[88:31] that I'm describing this like downward
[88:33] slope is basically like
[88:36] um like the weapon systems were reduced
[88:38] for Oblivion. The the you'd be you might
[88:41] be surprised having played Oblivion, but
[88:42] the magic system was was reduced for
[88:45] Oblivion. Um things just got simpler and
[88:47] simpler and simpler. Oblivion was a
[88:49] fantastic game. I think I think in a lot
[88:51] of respects you can say it was better
[88:52] than Morwind but um
[88:56] yeah under understanding the climate in
[88:58] which Morwind released I I still think
[89:01] Morowind's better if that makes sense.
[89:03] Um
[89:06] anyways and and Skyrim was a like
[89:08] massive simplification of things as
[89:10] well. Um
[89:13] I don't know. I can understand people
[89:14] might not agree with that argument, but
[89:16] I do think especially beyond Skyrim,
[89:19] Fallout 4 and
[89:22] Fallout
[89:24] 76, I think is what it's called. Um, and
[89:29] Starfields and some of the mobile games
[89:31] that they did and things like that. It's
[89:35] been
[89:38] I don't know. I still don't think Elder
[89:40] Scrolls 6 and I I talked big game about
[89:42] how I thought Starfield was going to be
[89:43] great. So, hey, maybe I'm going to be
[89:45] terribly wrong again. Um, but I still
[89:48] don't think
[89:48] >> Can you still call yourself a Bethesda
[89:50] fan?
[89:57] Do you can you be fully rational and
[90:01] logical about Bethesda or will you still
[90:04] have naive hope in spite of all evidence
[90:07] from the last 10 years
[90:11] cuz I think that's the line right is is
[90:13] >> that's tough though cuz like I did enjoy
[90:15] Fallout 4. Um, I thought it was a lot
[90:20] weaker than New Vegas and I thought it
[90:22] was weaker than three again.
[90:24] >> So, note that I said 10 years that it
[90:26] launched,
[90:27] >> Luke, I said 10 years.
[90:28] >> I'm trying to remember.
[90:29] >> Fallout 4 came out 11 years ago.
[90:33] >> Jesus.
[90:34] >> No, it didn't.
[90:35] >> Yes, it
[90:37] >> actually 10 and a half years.
[90:38] >> No, it didn't. Shut up.
[90:41] >> Holy crap. That
[90:43] >> I just broke him. I broke him down. I'm
[90:46] sorry.
[90:48] >> You want to know something funny? You
[90:49] know how I've like taken a break from
[90:51] streaming for a long time?
[90:53] >> I went through my most recent streaming
[90:56] assets folder.
[90:58] It's all Fallout 4 stuff. My taking a
[91:02] break has been a decade. Oh my god.
[91:06] Oh my god.
[91:07] >> He's dying, Dan. Holy crap.
[91:14] Okay. Um,
[91:17] >> oh man. Should I break him again?
[91:21] >> I How old is Mororrowind?
[91:23] >> Oblivion. I know the Xbox.
[91:24] >> Oblivion was 20 years ago.
[91:27] >> Yeah. Yeah. Yeah. Cuz Xbox 360 like I
[91:29] knew that one. For some reason, the gap
[91:31] between
[91:32] >> Oblivion and Fallout 4 seems so much
[91:35] bigger than between Fallout 4 and now. I
[91:37] don't know why that's true, but it feels
[91:39] like it.
[91:40] >> Um, okay. So the the question of the
[91:43] last 10 years thing is is tough.
[91:46] >> Yeah, it's tough.
[91:48] >> Even even if we look like locally,
[91:50] right? If we look at Landis Media Group,
[91:52] like Landis Media Group looking five
[91:55] years back is so much different in what
[91:57] we're doing, who works here,
[91:59] >> our capabilities are, all that kind so
[92:01] different. Go back go back seven years,
[92:05] >> completely different thing. Um,
[92:08] so it's okay. It's been an incredibly
[92:12] long time since Skyrim, but the only
[92:14] mainline game that they've released in
[92:16] that time. I don't I don't necessarily
[92:20] think I believe 76 is a mainline game.
[92:24] Is that fair to say?
[92:29] I don't know.
[92:31] >> This was meant to be kind of a yes or no
[92:33] answer Luke.
[92:34] >> Yeah. Sorry.
[92:35] >> Are you still a fan of Bethesda?
[92:37] Well, I I what I think I'm saying is I
[92:39] can't conclusively say either way
[92:42] because they haven't really released
[92:44] anything in the series that I liked in
[92:46] that period of time. They haven't
[92:47] released a mainline Fallout game. They
[92:49] haven't released a mainline Elder
[92:51] Scrolls game. Those are the two games
[92:52] that made me love Bethesda. So, I don't
[92:54] think I can say
[92:56] >> I I'm I'm going to still naively hold
[92:58] out hope though. Okay, then it sounds
[93:00] like you can still be at least a little
[93:02] bit of a fan
[93:03] >> cuz if you can naively hold out hope for
[93:06] something. I think
[93:07] >> I am reviews and buying ES6 day one
[93:10] guaranteed cuz I I have to experience
[93:13] it. I must there's no other like
[93:16] >> I think that's the answer. I think
[93:18] that's the answer then.
[93:20] >> Um I want to jump into a couple sponsor
[93:23] spots so that uh because I'm sure you
[93:24] want to get on with your day at some
[93:26] point. Uh, the show is brought to you
[93:27] today by ODO. Having to cycle between a
[93:30] dozen different tabs and apps when
[93:32] operating a business can have you
[93:33] wanting to jump into a bottomless pit of
[93:35] despair, but not with our sponsor, ODU.
[93:38] ODU is the all-in-one business
[93:39] management software that has just about
[93:41] any app your business could need
[93:42] integrated into one platform. You and
[93:45] your team can get projects off the
[93:46] ground with ODO's designated app, so
[93:48] everyone can stay organized and on the
[93:50] same page. Once your project gets moving
[93:52] along, their CRM tool can help you send
[93:54] quotes and schedule meetings in a simple
[93:56] drag and drop interface. And the
[93:58] invoicing app can help you get paid when
[94:00] the time comes with things like sending
[94:02] invoices and tracking open accounts. And
[94:04] they can be as automated as you'd like.
[94:06] Something that makes ODO even cooler is
[94:08] that if you only need a single app, you
[94:10] can just use it for free. Use our link
[94:14] for a 15-day trial with no credit card
[94:16] required or book a demo with their
[94:18] expert team to learn how ODU can help
[94:20] your business grow. The show is also
[94:22] brought to you by Op manager Nexus. Op
[94:25] manager Nexus is a full stack
[94:26] observability platform that gives you
[94:28] and your IT team a single clear view of
[94:31] your entire infrastructure. There's no
[94:33] need to start from scratch. It
[94:34] seamlessly plugs into the manage engine
[94:36] ecosystem and the tools that your team
[94:38] already relies on. Whether your org
[94:40] works from the cloud or prefers an
[94:42] on-remise setup or a mix of the two, Op
[94:44] manager is designed to work the way you
[94:46] prefer. It's built to be deployed
[94:48] quickly and uses customizable dashboards
[94:50] so each member of your team gets the
[94:52] reports and information they need the
[94:53] most. Plus, with a global network of
[94:55] localized data centers, everything stays
[94:58] highly available because they know just
[95:00] how much reliability matters. It's
[95:02] backed with strong data encryption
[95:04] adhering to industry and regional
[95:06] compliance rules so your data stays
[95:07] protected. Op manager Nexus, stop
[95:09] guessing and start knowing with full
[95:11] stack visibility. We'll have a link for
[95:13] you down in the video description.
[95:15] >> All right, Luke, want to pick a topic or
[95:17] two?
[95:19] >> Uh, yeah, sorry. I was talking to Chad
[95:21] about Bethesda things.
[95:22] >> Oh my god.
[95:22] >> Um,
[95:23] >> definitely fan. Still a fan.
[95:25] >> You got me started, bro. You got me
[95:26] started. It's not my fault. Um,
[95:29] >> yeah, let's do this one. We were talking
[95:32] about AI stuff before we did the the
[95:35] jump off topic section, so we'll go back
[95:37] to it for some reason. Um, New York
[95:40] State could ban data center development.
[95:44] New York State has put a temporary hold
[95:46] on any further data center development.
[95:48] Governor Kathy
[95:51] Ochel uh will have the final say and
[95:55] must decide before December if the state
[95:57] will enact a permanent ban on all data
[96:00] centers that demand 20 megawws or more
[96:03] of power. Uh as as a quote uh this is
[96:06] one of the first times that we're really
[96:08] drawing a line in the sand and saying
[96:10] that as a state legislature we have the
[96:13] responsibility to make sure New Yorkers
[96:15] are in the driver's seat. uh said New
[96:18] York Senator Kristen Gonzalez. And then
[96:21] again to quote, "Big tech has been used
[96:25] uh has been used to writing their own
[96:26] rules or not having rules that they have
[96:29] to play by, which is yeah, I mean, fair
[96:31] enough." Um
[96:36] I feel like Yeah, it's interesting. I
[96:38] wonder what the
[96:42] >> just for context a 20 megawatt data
[96:44] center would be approximately 133 SpaceX
[96:49] AI satellite data centers.
[96:53] >> H
[96:55] just throwing that out there.
[96:59] >> That is pretty fun for context actually.
[97:02] Um,
[97:03] >> in case anyone was wondering why some
[97:07] people are pushing back on the idea of
[97:09] AI satellite data centers, um, it would
[97:13] take 133 of those to build something
[97:16] that's big enough that the state of New
[97:17] York, it wants to ban it. Wants to ban
[97:20] it.
[97:22] Not the orbit of New York. Um,
[97:27] yeah. Interesting. I'm trying to rack my
[97:30] brain. What are the primary complaints
[97:33] about data centers being near you?
[97:35] There's power obviously. There's water
[97:38] >> that it's pulling out.
[97:40] >> Ugly.
[97:40] >> Noise. I guess some people have
[97:42] mentioned the really high pitch noise as
[97:43] well.
[97:44] >> Ugly. Y
[97:45] >> I'm sure there's like ecological
[97:47] destruction
[97:47] >> cost. Uh so that's related to the power,
[97:50] but also
[97:51] >> the power cost thing. The power cost
[97:53] thing feels solvable
[97:55] >> um
[97:55] >> by making the power for the data center.
[97:58] I think part of it too much is like the
[98:01] economic argument for data centers is
[98:03] all the the jobs that they create, but
[98:05] that's only while they're building them.
[98:07] >> Um, once they're built,
[98:09] >> yeah, they don't have much afterwards.
[98:10] >> Modern data centers require very little
[98:12] in the way of personnel compared to
[98:14] while they are being constructed.
[98:17] >> Um,
[98:17] >> totally. So, it's one of those things
[98:19] where when they talk about all the jobs
[98:20] that like AI infrastructure is creating,
[98:22] it's like, yeah, as long as we're
[98:25] chasing infinite growth, then it will,
[98:27] you know, continue to support those
[98:30] jobs. But the second that there's a
[98:32] cutback or bubble burst or the the
[98:34] buildout is stalled, then they kind of
[98:36] go away.
[98:42] >> And it's it's worth noting actually. I
[98:44] think there's it's man it's it's kind of
[98:48] funny to kind of I I don't know I don't
[98:51] have uh I don't think I've taken a
[98:53] strong position either way on the the
[98:55] building out of data centers and it's
[98:58] really it's really interesting to see
[99:00] both the good points that that both
[99:02] sides of this issue make and also the
[99:05] somewhat seemingly flawed points that
[99:08] both sides make. Like there's uh there's
[99:10] a lot of talk about the water
[99:12] consumption of data centers. Um and
[99:15] during construction that is that is like
[99:18] super true. They use a ton of water. Um
[99:22] but a big part of that is that it takes
[99:24] a ton of water to like make concrete
[99:27] construction. Um
[99:30] like back when um uh back when I toured
[99:33] u Intel's FAB, they were they were
[99:35] showing me the site where they were they
[99:38] were building like an extension to the
[99:39] fab and they literally like this the
[99:42] scale of these kinds of projects. They
[99:44] had built their own concrete production
[99:46] facility on site so that they could so
[99:50] they could mix and then logistically
[99:54] move around the concrete much more
[99:56] efficiently. And like people don't
[99:58] necessarily recognize like how much
[100:01] water that consumes, but once it's
[100:05] built, as long as they're not using
[100:06] evaporative cooling, which from my
[100:08] understanding is kind of going out of
[100:10] fashion. Uh
[100:12] >> my understanding, a lot of these data
[100:13] centers are using evaporative cooling.
[100:16] >> The newer stuff seems to be getting away
[100:19] from it, though. And so if they're not
[100:21] using evaporative cooling, then their
[100:24] ongoing water consumption is actually
[100:26] pretty low. Um,
[100:30] yeah, Cphilus says, "Building a concrete
[100:32] production facility is common in a lot
[100:34] of construction." Yeah, anytime you're
[100:35] doing a large scale project, it's like,
[100:37] gee, I don't know. Should I truck the
[100:39] concrete from there to here um in like
[100:42] cement mixers or should I just make
[100:45] everything here out of the raw
[100:47] materials?
[100:52] Uh Jlock said, "Nobody would have an
[100:54] issue with the power issues, water
[100:56] consumption, even noise, etc. if data
[100:57] centers are being used to solve medical
[100:59] mysteries, world hunger, major
[101:00] infrastructure engineering problems,
[101:02] etc., rather than for a few billionaires
[101:04] to spy on us and get even more rich."
[101:06] You are forgetting um funny AI cat
[101:10] videos though.
[101:12] >> Yeah. and also forgetting that a lot of
[101:15] people are using it to do legitimately
[101:18] useful things. Um,
[101:19] >> yep,
[101:21] >> including unfortunately Palanteer. Um,
[101:24] >> yeah, we talked about that early in the
[101:26] show. Palunteer again, classic stop
[101:28] clock is right twice a day. Palanteer
[101:30] did something good, but also scary
[101:32] because it's Palanteer.
[101:36] >> Porto says, "All is for good. All is
[101:38] forgiven. Cute cats."
[101:40] I don't know, man. I um it's always
[101:43] interesting for me to get boots on the
[101:46] real ground, you know, like talk to
[101:48] people who are outside of my bubble and
[101:50] especially don't take this the wrong
[101:52] way, Wan show viewers, like you guys are
[101:55] you guys are like dialed. You're the
[101:56] you're the LTT core, right? Um and it's
[102:00] always interesting for me to go outside
[102:01] of that bubble and talk to people about
[102:03] this stuff. I was playing badminton in
[102:04] Taiwan. Stop me if I told this story
[102:07] last week on Wow. Um, and I was just
[102:10] chatting with um with one of the one of
[102:12] the players at at the club and um he
[102:15] comes up to me. His English is pretty
[102:17] good and uh we've we've chatted a couple
[102:19] times over the years cuz I come back
[102:20] every year. So people tend to kind of
[102:23] they're like, "Oh yeah, that that white
[102:24] guy, right?" Um yeah, yeah, hi. It's me.
[102:26] Um but at some point he kind of put two
[102:30] and two together and he started like
[102:31] watching the channel. So now when I
[102:32] showed up he's like oh you know hey
[102:34] Lionus uh and he asked me he's like what
[102:36] are your thoughts on AI? And I'm like,
[102:37] well, that's a very broad um question.
[102:40] Um, you know, I think we're both here to
[102:43] play badminton, not talk about this for
[102:44] the next 3 hours, but um I'm going to
[102:47] uno reverse this. What do you think
[102:50] about AI? And he kind of goes, oh, well,
[102:53] like again, uh that's a very broad
[102:56] question. What specifically do you mean?
[102:57] And I'm like, well, like what do you do
[102:59] for work? You know, what do what do you
[103:02] use it for? What does your company use
[103:04] it for? you know, like someone boots on
[103:06] the ground just
[103:08] out there living their life, working,
[103:11] eating, sleeping, playing badminton. Um,
[103:14] I wanted to hear his perspective. And
[103:16] obviously, this is a sample size of one.
[103:20] But his response was really interesting
[103:22] to me. He's like, "Well, I I work in
[103:24] like data compliance." And I kind of go,
[103:26] "Oh, okay." You know, like what does
[103:28] that mean exactly? He's like, "Well, we
[103:30] provide services for things like um law
[103:33] offices and medical clinics and things
[103:35] like that." And I'm like, "Okay, yeah,
[103:36] that that gives me a pretty good idea of
[103:38] what you guys are up to." And um I'm
[103:41] like, "Okay, so like how would you say
[103:43] that, you know, AI has changed the way
[103:45] that you guys operate?" And he's like,
[103:47] "Well, we have about the same headcount
[103:49] as we used to, but our marketing
[103:52] department is completely gone. Um we
[103:56] deploy tons of code to production. that
[103:59] is overseen
[104:01] um but very much generated code. It has
[104:06] made us wildly more productive. So we
[104:10] have about the same headcount that we
[104:11] did before we started using it but our
[104:14] revenue and productivity is like
[104:16] massively up and um where where I kind
[104:20] of like I wanted to poke and prod at
[104:22] that a little bit right and I was like
[104:23] okay you know what does this mean for
[104:25] the maintainility
[104:26] >> measuring that too
[104:28] >> of these projects and like so you know
[104:30] oh yeah how do you measure that well you
[104:31] measure it in revenue you measure it in
[104:32] completed projects apparently it's like
[104:35] >> that's that's reasonable yeah I just
[104:37] yeah
[104:38] >> and uh and in terms of like like
[104:40] maintenance. He goes, you know, well,
[104:41] we're real software engineers, so like
[104:44] you know, we can do that. Um, but using
[104:48] this to like turbocharge our
[104:49] productivity and, you know, completely
[104:52] just eliminate entire departments
[104:56] um has been has been something that has
[104:59] been very I'm going to use finger quotes
[105:01] here, successful for them. And that
[105:06] perspective was like I said before, it
[105:08] was a sample size of one, right? So
[105:10] let's not treat it as anything deeper
[105:12] than what it is. But when I hear people
[105:16] say, and and this is this is a really
[105:18] common sentiment that shows up in our
[105:19] comments section, in our chat, wherever,
[105:21] that nobody wants this, I kind of go,
[105:25] okay, but that's not actually true.
[105:30] And I think we could argue all day
[105:33] about, you know, where on the spectrum
[105:35] the people who want it and appreciate it
[105:38] and enjoy it and use it and are and are
[105:40] benefiting from it versus the people who
[105:43] adamantly oppose it and and won't touch
[105:45] it. Like we we could kind of we can
[105:47] argue all day about sort of where that
[105:49] split is, right? Um but what is
[105:52] objectively not true is that either
[105:55] everyone is benefiting from it and they
[105:57] need to just get on board. That's not
[105:59] true. or nobody wants this. It's it's
[106:02] it's purely it's purely negative and
[106:05] nobody could benefit. That's also not
[106:07] true. And so I just am I'm still taking
[106:10] a very
[106:12] wait and see approach. Um we use it for
[106:17] some things internally. We've
[106:20] experimented a little bit more with
[106:22] using it externally. Um, I think that
[106:26] there's absolutely a value to
[106:29] specifically calling out when something
[106:31] is not AI like this. This mouse pad
[106:36] would have functionally no artistic
[106:40] value to me if a human hadn't drawn it.
[106:44] That doesn't mean it would be worthless.
[106:46] I think that someone probably would
[106:48] enjoy an AI generated desk pad if it was
[106:51] like, you know, cool and it they just
[106:53] liked it. They vibed with it. But for
[106:56] me, like this has has has an artistic
[107:00] value that is
[107:02] it's just hard to quantify. How do you
[107:04] quantify that, Luke? Can you quantify
[107:05] it?
[107:08] >> And how do you properly prove it? Um,
[107:12] >> oh that's sad.
[107:13] >> I think this is this is
[107:14] >> we've talked about this before
[107:16] >> like how cottage industry and like Etsy
[107:18] is going to have to move to like selling
[107:20] the live stream of you creating the art,
[107:22] you know,
[107:23] >> and then how do you prove that the live
[107:25] stream isn't AI generated as well. Um,
[107:29] then that that whole series gets gets
[107:31] really weird. But like something
[107:32] something I'm kind of running into is
[107:34] there was a game recently that said
[107:38] I don't know I think I think
[107:42] I think they said that they're using AI
[107:44] stuff for ideiation, but they're not
[107:46] going to have it in their finished game.
[107:48] And then they had some Oh man, this
[107:50] isn't a topic of the doc. I'm trying to
[107:52] go off the top of my head. They had some
[107:53] like playable demo
[107:56] and people found AI assets in the
[107:58] playable demo and the company's
[108:00] basically like, "Oh, like that that
[108:03] wasn't supposed to be there. We didn't
[108:05] like realize that was there. um we're
[108:08] going to try to figure out what happened
[108:09] or whatever. I I think employers
[108:14] aren't going to be able to even know if
[108:17] their employees are really using it or
[108:19] not.
[108:19] >> No.
[108:20] >> Um
[108:21] >> I wouldn't know if Dan was an AI right
[108:23] now.
[108:31] >> No response. Didn't move at all. Yeah,
[108:33] that would be pretty hard to tell. Uh,
[108:35] >> actually, I think an AI might have had
[108:36] less latency on it. So, I think he's
[108:38] probably a human.
[108:40] >> I'm not an AI, Lionus.
[108:42] >> That's actually a pretty damn good AI
[108:44] voice.
[108:46] >> Now, you got to do the like you're out
[108:47] of
[108:47] >> You can't ask me to do that. I can't uh
[108:50] I can't confirm or deny
[108:52] >> small language model.
[108:56] >> Um, but but yeah, it's like even if it's
[108:59] just for ideation, you can't necessarily
[109:01] control it. someone maybe they're on
[109:03] their lunch break and they're stuck on
[109:04] an issue and they they want to look good
[109:06] for their boss so they u AI brainstorm
[109:09] some junk on their phone like you can't
[109:11] um
[109:13] >> and it's funny even
[109:16] >> even very anti- AI people um are it's
[109:21] it's it's really interesting to watch
[109:22] the conversations about it because like
[109:25] we've experimented with using a little
[109:26] bit of AI generated stuff in our
[109:29] thumbnails recently
[109:30] >> I've seen that
[109:31] >> um This one was was sort of universally
[109:36] disliked for being uh AI AI core. Uh
[109:41] except that I actually this is an actual
[109:43] photo of me sitting on that manhole
[109:46] cover. Uh I actually don't know what Sam
[109:49] did with these uh with the with the uh
[109:51] the wires. Those weren't there. I'm not
[109:53] sure if he generated them or if he like
[109:56] put some other asset in there. Um,
[109:58] >> that sign wasn't there either.
[110:00] >> But, but what's funny is as much as
[110:03] people hated those wires,
[110:06] one person, as far as I can tell, one
[110:11] person noticed this.
[110:16] This has been up for a month.
[110:19] That's not me.
[110:22] >> I noticed that.
[110:23] >> Then two people did. my
[110:25] >> I'm not just trying to like be like,
[110:27] >> "Yeah, that's fine." My own wife, I
[110:29] asked her about this thumbnail and
[110:31] without prompting her, without like
[110:33] leading her, she basically just started
[110:36] talking about, "Oh, yeah, I I think
[110:38] it's, you know, the the monitor is cool
[110:40] or, you know, whatever, but didn't
[110:43] notice that I only have one earring in
[110:45] my left ear." Um, there's a Oh, didn't
[110:49] notice I didn't have braces.
[110:51] >> There's a couple little problems
[110:52] >> with your eyes. I don't know what it is,
[110:54] but there's something going on with your
[110:55] eyes that tells me something's off.
[110:56] >> Uh Sammy Sammy noticed it and was like,
[110:59] "Hey, what's up with this?" And I
[111:00] basically was like, "Yeah, it was an
[111:01] interesting it was an interesting
[111:04] experiment. A human still worked on it."
[111:06] So,
[111:06] >> you want to see something? Sorry.
[111:08] >> The AI Linus is AI, but actually the uh
[111:12] and the the cardboard is AI, but I
[111:15] believe the monitor um was um Sam's work
[111:20] and then the background is Sam's work.
[111:22] Um, so a human still like composited
[111:24] everything,
[111:25] >> but that's not that's not me. And very
[111:29] very few people noticed, which I thought
[111:32] was really interesting. That's it was I
[111:35] had a long conversation with Sammy about
[111:36] it cuz he's like, "Look, I like is this
[111:38] our direction? Are we just going to stop
[111:40] taking pictures of you and just just
[111:42] generate you?" I'm like, "No." Uh, every
[111:44] video that we've done since then is a
[111:46] real is a real picture of me for for
[111:48] better or or for worse, including this
[111:51] was another one uh that people were
[111:53] really upset about because a Gemini
[111:56] watermark was on the original version of
[111:58] it because that PC case was added with
[112:01] AI, but that was the only thing AI about
[112:04] it. That is otherwise just a photograph
[112:06] of me in the uh in the PC build corner.
[112:09] So, it was just the case that was thrown
[112:11] on there. So, What's funny is like
[112:16] the things that people thought were AI
[112:19] about this were not the things
[112:21] necessarily that were actually AI about
[112:24] it. The the the watermark made some
[112:27] people pretty upset. But then we had one
[112:30] where it was literally completely AI
[112:32] generated Lionus and almost nobody
[112:34] noticed because the watermark wasn't
[112:35] there.
[112:36] What's funny about that one and I've
[112:38] I've probably just seen your freaking
[112:40] face too much, so I recognized it. But
[112:43] cuz I I didn't actually I didn't clue
[112:45] into it based on the cardboard sign or
[112:48] the Linux logo or anything else. It was
[112:50] just I looked at your eyes and my brain
[112:52] was like
[112:53] >> they're not quite the right color.
[112:55] >> So was that I I don't even know what it
[112:57] was. Just something in my head was like,
[112:59] ah, that's something's wrong. It's like
[113:01] not him. Um, and I don't know if the I I
[113:05] I wasn't necessarily aware if maybe the
[113:07] editors were just trying to like do
[113:09] touch-ups and went a little bit too far
[113:10] or something like that. Uh, but but for
[113:13] me, my brain was like, "Yeah, it looks
[113:14] like
[113:15] >> so
[113:15] >> reminds me of AI generated."
[113:17] >> So, that's a fun thing. Uh, I've
[113:18] actually had that conversation with our
[113:20] thumbnail artist because I found that
[113:22] some of our real thumbnails actually do
[113:25] look a little bit AI. This one's not.
[113:28] This is a real picture, but it has the
[113:30] same eye lightning that might trigger
[113:34] the same reflexive response from you as
[113:36] the eyes in the other one, if I had to
[113:38] guess.
[113:38] >> No. Can Can you Can you like screenshot
[113:40] one of these and then scroll to the
[113:43] other one so we can have both on screen
[113:45] at the same time?
[113:45] >> I think we can just That one doesn't
[113:48] make me think that it's off. Like I I
[113:51] don't know what it is. There's some
[113:53] there's something about the eyes in that
[113:54] AI one that
[113:56] >> I thought this one looked really AI.
[113:59] Personally,
[114:00] >> I I thought like the the the crayon in
[114:03] your hand gives me or whatever that is,
[114:05] the marker in your hand gives me vibes
[114:07] similar to the the cardboard on fire
[114:10] thing. Um
[114:12] >> so that that might have made me think
[114:13] something was going on there. But yeah.
[114:16] Um you want to see something on the
[114:17] other end of the spectrum here real
[114:18] quick?
[114:19] >> Sure. Yeah. uh a content creator I've
[114:21] been enjoying watching for quite a while
[114:23] now as some like real solid second
[114:26] monitor content. Um but yeah, bring it
[114:29] up on YouTube. No, I don't think that's
[114:30] a dig. I think that's part of the goal.
[114:32] >> Um and I'll I'll I'll I'll you know,
[114:35] I'll lock in for interesting segments,
[114:37] but uh look up Gotham Chess
[114:40] and look at his thumbnails
[114:43] and show those on screen.
[114:47] Okay, so this is fully embraced.
[114:51] >> Click on the video tab.
[114:52] >> The AI.
[114:54] Oh my god.
[114:56] >> Um,
[115:00] and
[115:02] it feels like there I I don't know what
[115:04] it is because it's been really
[115:06] interesting to me since he started doing
[115:08] this. Um, actually checking the comments
[115:11] and and seeing what the sentiment on it
[115:13] is. Um, you guys are usually second
[115:16] monitor content for me. I don't take
[115:18] that as a dig. What? W show is like
[115:20] supposed to be second monitor content.
[115:22] Like that's
[115:23] >> that's crazy. Uh, sorry. Alex Badminton
[115:26] Academy, who I did a collab with a
[115:28] little while ago, just uploaded a collab
[115:29] with TechJoice. That's cool.
[115:31] >> Oh, nice. That is cool.
[115:33] >> Anyway, he's using AI AF in
[115:36] >> He's clearly using Oh, this guy is
[115:38] >> Yeah.
[115:39] >> Oh, interesting. So, this was another
[115:41] example that I wanted to show, but um
[115:45] >> here we go.
[115:45] >> This is interesting because it's it I I
[115:48] thought it might have been
[115:51] >> See, that doesn't even necess that just
[115:53] looks like he might have done like weird
[115:55] head enlargement or something. It
[115:56] doesn't necessarily
[115:58] >> immediately scream out AI to me. Um the
[116:00] Gotham Chess one, the commenters seem to
[116:03] be having a lot of fun with it. Um, like
[116:05] I I I can pretty much guarantee you if
[116:07] you go to the Gotham Chess video of him
[116:09] as a sumo wrestler and check the
[116:11] comments that there will be comments
[116:13] like talking about the thumbnail,
[116:16] >> but in a either neutral but surprised
[116:19] way or in a like supportive this has
[116:23] been kind of fun. His like feray into
[116:25] crazy AI thumbnails way.
[116:26] >> Here's another completely generated
[116:28] Linus. That is not my racket. Those are
[116:30] not my shoes. I was not wearing shorts.
[116:33] your head shape is wrong.
[116:35] >> Yeah. So, it's the it's the same here.
[116:37] This is completely generated.
[116:39] >> Um
[116:40] >> Yeah.
[116:41] >> So, some people are just leaning right
[116:42] into it and no one cares slash it's it's
[116:46] neutral and then others are using it a
[116:48] little bit and facing a bunch of
[116:49] backlash. I saw Jay's two cents was
[116:51] getting lit up on Twitter about using AI
[116:53] on something. I actually haven't looked
[116:54] at his thumbnails recently. The thing
[116:56] that I'm stuck on is I can't tell if
[116:58] it's a creator difference, an audience
[117:00] difference, or a style difference.
[117:03] Because with Gotham Chess, there's no
[117:04] real question.
[117:06] >> It doesn't seem like he's trying to get
[117:07] away with using AI to generate his
[117:11] thumbnails.
[117:12] >> He's playing with it very openly. It's
[117:15] very obvious. He's not trying to hide
[117:17] it. Um, it's it's it's kind of fun. Um,
[117:22] and then the the audience seems to be
[117:24] vibing with it. It seems to me, and I'm
[117:28] I am jumping to conclusions here. I
[117:31] don't have a lot of evidence for this,
[117:32] but I I if I had to wager right now, it
[117:35] seems to me like the thing that's
[117:37] bothering people is when it seems like
[117:42] I don't know, too normal,
[117:44] >> like when they're trying to like in the
[117:45] uncanny valley.
[117:47] >> I mean,
[117:47] >> I guess. Yeah.
[117:49] >> Pretty well doumented.
[117:52] >> Um Yeah. Yeah. Yeah.
[117:55] Yeah, it's interesting. It's uh just
[117:58] seeing seeing where the community is
[117:59] drawing the line because there there's
[118:00] clearly creators that people are fine
[118:03] with it.
[118:04] >> As uh SJ John Trombbley says, I think
[118:07] it's a community difference. I was
[118:08] talking to a VTuber a few weeks ago and
[118:10] she said you basically get
[118:12] excommunicated if you use AI for
[118:14] anything.
[118:16] which is interesting because
[118:19] I could see that being a space where it
[118:22] would be used a lot um
[118:26] >> in in some period of time from now uh
[118:30] when that when that sentiment when when
[118:33] someone is the first one to use it a
[118:35] bunch and and be successful anyway. Um
[118:51] It's a It's a It's such an interesting
[118:54] realm.
[118:57] Oh, someone says already happened.
[118:59] Looked up Val 987.
[119:02] Are they just like Aia AF?
[119:06] I don't even know what I'm looking at
[119:08] here.
[119:10] >> V E D A L It looks like. Yeah. I can't
[119:14] uh I can't tell what this is.
[119:18] Neurosama. Is that it?
[119:20] >> I have no idea.
[119:21] >> Yeah. I don't think I'm going to be able
[119:23] to figure this out.
[119:31] Okay.
[119:35] Yeah. I I think the AI Vtuber. Okay.
[119:38] Yeah. I I this is a this is a deeper
[119:42] rabbit hole than I think I'm going to
[119:43] dive into
[119:45] live on show. Thank you for flagging
[119:48] that for me though.
[119:50] >> Uh why don't we jump into another topic
[119:52] here? I don't even remember how we got
[119:53] on the subject.
[119:55] >> Uh oh yeah, the New York State could ban
[119:57] data center development. All right,
[119:58] cool. Uh hey, here's a cool thing that
[120:02] some YouTuber
[120:06] slash did. Uh, Zerof put up an
[120:10] incredible
[120:11] >> incredible video interviewing a bunch of
[120:14] experts and diving into I'm not going to
[120:18] play the video cuz you guys need to go
[120:20] watch it. It's 45 minutes, but it is a
[120:23] it is a hard and fast 45 minutes. Um, I
[120:28] can give you guys some of the some of
[120:30] the the top level details, but it's
[120:32] worth watching just so you can dive
[120:35] right into it. Um, it's an amazing
[120:38] investigation into Finnish startup
[120:40] Doughut Lab, who made waves at CES 2026,
[120:44] claiming to have a revolutionary solid
[120:46] state sodium ion battery with 400 W
[120:50] hours per kilogram of energy density, a
[120:52] 100,000 cycle life, and 5 minute
[120:55] charging. So, put that in context. uh
[120:58] 400 W hours per kilogram, which would be
[121:02] like 33% higher than a pretty solid
[121:06] lithium ion cell. 100,000
[121:10] cycle life is functionally a forever
[121:13] cell. Like whatever device you put that
[121:16] battery in would be dead long before you
[121:18] would reach a 100,000 cycles. And 5
[121:21] minute charging
[121:22] >> functionally useless compared to
[121:24] anything else.
[121:24] >> Yeah, 5 minute charging. That's nuts.
[121:27] That basically solves it. I mean, I
[121:29] finally Oh, hey, Luke. I finally got my
[121:31] motorbike.
[121:33] >> Yeah, Sammy showed me some pictures,
[121:37] >> man. That looks freaking sweet.
[121:39] Actually,
[121:41] >> yeah, I haven't shown the W show yet.
[121:44] >> Uh, wait, how is Dan so far down in my
[121:47] stupid list? Hello,
[121:50] >> Besser.
[121:51] There he is. All right, I'll send these
[121:53] over and we can fire these up here. Uh,
[122:00] Oh, here I'll also fire you over the
[122:02] picture of discovering uh why my GPU
[122:06] never showed up in my server um after
[122:09] all this time. And I finally got around
[122:12] to opening it up to troubleshoot the uh
[122:14] the problem. Uh okay. So, let's uh let's
[122:17] let's do motorbike picture first.
[122:20] So, I 5 minute charging. I had to I had
[122:23] to fuel up a vehicle for the first time
[122:25] in a while because my car is electric.
[122:27] And so I got my I had my motorbike back
[122:29] and it probably took me like five
[122:30] minutes because I went to that gas
[122:32] station uh you know the one that you
[122:34] have the loyalty to that doesn't have
[122:36] pay at the pump.
[122:37] >> Yeah.
[122:37] >> Up over hill. Yeah.
[122:38] >> I stopped by there to to grab some gas
[122:40] and so I like I had to park and then I
[122:44] had to go inside. I had to leave my
[122:46] credit card with them which didn't make
[122:47] me deeply comfortable. Uh but they take
[122:50] your credit card as collateral. then you
[122:51] go put as much gas in the tank as you
[122:53] want. I had no idea how much fuel I
[122:55] needed because it's been 3 years since I
[122:57] fueled my motorbike. One, and two, the
[123:01] cost of fuel is so different from the
[123:02] last time I fled it that even if I did
[123:04] remember how much money I t because I
[123:06] don't remember it in liters, right? So I
[123:08] >> got me.
[123:09] >> Yeah. I So I had no idea. I was able to
[123:13] super accurately guess
[123:16] for a really long time and then uh some
[123:20] dude who said he wouldn't start wars or
[123:23] something started a bunch of wars
[123:26] something you know whatever uh and now
[123:28] the gas price is insane and then I can't
[123:30] guess anymore. But yeah, you don't have
[123:32] to this doesn't matter for the show but
[123:35] you don't have to leave your card. you
[123:36] can you can just um say a certain amount
[123:39] and then go back in and they'll refund
[123:41] it.
[123:42] >> But I didn't know. So I just I was like,
[123:44] "Okay, sure. Here, take my credit card.
[123:45] I'll put in as much as I can and then
[123:47] I'll come back and get it." So it
[123:48] probably took me about 5 minutes to then
[123:51] go and fill it and then go and do the
[123:52] transaction and then go like get back on
[123:54] my bike. So if you could if you could
[123:57] refuel an electric vehicle in 5 minutes,
[124:00] that's functionally in my experience
[124:05] so similar to fueling a gasoline car
[124:08] that it might as well be the same thing.
[124:10] A new investigation led by battery
[124:13] researcher Zeroth, pulling in over 20
[124:16] independent experts, has now concluded
[124:18] that the miracle battery is literally
[124:21] just a standard lithium ion cell. You'll
[124:25] have to watch the video, but here's a
[124:27] summary of the findings. First up is the
[124:29] voltage curve. The cell sits at 3.7 to
[124:32] 3.8 volts at 50% charge. That is exactly
[124:36] where high nickel lithium ion cells
[124:38] operate. A real sodium ion cell doesn't
[124:40] go much past 3 1/2 volts. Second, and
[124:43] more damning, is the cell expansion
[124:45] data. When a battery charges, ions
[124:48] squeeze into the anode and make it swell
[124:50] in a predictable pattern. A graphite
[124:52] anode produces a distinctive kink in
[124:54] that expansion curve around 50 to 70%
[124:56] charge. The doughut cell shows exactly
[125:00] that kink, which is impossible for
[125:02] sodium ion because sodium ions are
[125:04] physically too large to fit into
[125:05] graphite.
[125:06] >> Don't shame it. Researchers compared the
[125:09] combined evidence to having both a
[125:11] fingerprint and a photo of the suspect.
[125:14] Real world capacity also came in at
[125:16] about 298 W hours per kg, which is
[125:19] decent for lithium ion, but nowhere near
[125:21] the 400 that was claimed. Also, the
[125:24] structure behind this thing is a mess of
[125:26] shell companies. The tech traces back to
[125:28] a German firm called CT Coatings, whose
[125:32] patents include things like screen
[125:34] printed paving slabs and warning
[125:36] triangles.
[125:38] Nordic Nano was supposedly manufacturing
[125:40] the cells, but reportedly never made a
[125:42] single battery. Donut Lab handled the
[125:44] selling. The company raised around $25
[125:47] million from over 1300
[125:51] mostly small investors, which is really
[125:53] sad. These are regular people who put in
[125:56] anywhere from $3,000 to $23,000 each and
[125:59] then they inflated their valuation to
[126:01] $1.2 billion off of CES hype. Finnish
[126:06] financial and criminal authorities are
[126:08] now investigating. It's nice to hear
[126:10] that there are countries where
[126:12] misrepresenting your product is an
[126:13] actual crime when you take money from
[126:15] investors. That's pretty cool. Our
[126:18] discussion question is venture
[126:19] capitalists probably would have caught
[126:20] this with basic due diligence. I mean
[126:22] yes and no. Elizabeth Holmes, right? But
[126:25] whatever. So, the company went straight
[126:26] to small retail investors who couldn't
[126:28] do that. What does it say about
[126:29] crowdfunded tech that the people with
[126:31] the least protection are the ones left
[126:32] holding the bag? Usually, it doesn't
[126:35] work out this way, but boy, when it
[126:37] does, is it ever a giant boatload of
[126:39] suck?
[126:40] >> Isn't that sort of always?
[126:43] >> Um,
[126:44] >> okay. Let's go back to the I I I I I
[126:47] think I understand you were saying
[126:48] something different than the path I'm
[126:49] going down to be clear, but Okay. If we
[126:51] go back to like the NCX NCIX closure,
[126:55] like the banks and stuff get paid out
[126:57] first, right? Like the really big
[126:59] >> Yeah, it's a different thing like
[127:00] bankruptcy proceedings versus like like
[127:03] venture capital. But okay,
[127:05] >> a lot of the time the the the
[127:09] >> the the larger organizations have in
[127:12] both cases either the the the due
[127:16] diligence capacity or the legal clout to
[127:22] better protect themselves compared to
[127:24] small mom and pop investors. Yeah, it's
[127:27] a definite problem on both sides.
[127:30] >> Yeah.
[127:32] Anyway, super super disappointed to hear
[127:35] about this. I I never claimed to know
[127:38] enough about cell chemistry to have
[127:40] taken a strong stance one way or the
[127:42] other, which is um
[127:47] I'm which means that, you know, I was
[127:50] hopeful that it was a cool, awesome
[127:52] solid state battery thing, but I also
[127:54] wasn't like, you know, sure that it
[127:58] wasn't. Like, I just didn't really have
[127:59] a position on it. And now I'm uh
[128:02] >> that's also why it's cool that Zeroth is
[128:04] doing this.
[128:04] >> Glad to have closure. Yeah.
[128:06] >> Yeah. And and somebody who does know
[128:08] what they're what they're doing um is
[128:10] working on it. YouTube is such a cool
[128:12] place.
[128:13] >> And uh hey, I guess I'm ready to show my
[128:17] finished paint job.
[128:19] >> It's a It's an 8ft job, but I took these
[128:21] pictures from 8 ft away, so it should
[128:23] look pretty good. Throw it up there.
[128:26] >> So that's it. Damn, that's actually so
[128:29] sick.
[128:30] >> I saw them on Sammy's little phone, but
[128:32] they look way better full.
[128:34] >> Um, this is kind of in the This is in
[128:35] the shadow away from the sun. You can
[128:38] really see on the uh on the front wheel
[128:40] how there's kind of a bit of a color
[128:42] shift to it. So, yeah. Dan, do you want
[128:44] to go ahead and There she is. So, same
[128:47] here. You can kind of see how it looks a
[128:48] bit little bit This is more in the sun,
[128:49] so it looks a little lighter toward the
[128:51] back of the bike. This is just like
[128:53] pictures on my fold. Um, yeah, there she
[128:56] is. The shop did an amazing job of
[129:00] putting her back together. There's a lot
[129:02] of problems with the paint job, like I
[129:04] said. Um, not least of which is that
[129:07] there's some adhesion issues between the
[129:08] white paint layer and the primer that
[129:10] are going to make it not last that long.
[129:13] But for a 2003, boy, does she ever look
[129:17] pristine right now. New exhaust sounds
[129:20] amazing. I'm not normally like a like,
[129:22] oh man, yeah, you should like check out
[129:24] my exhaust. This sounds so good. Um, I'm
[129:27] not really that type, but
[129:30] um, they managed to do what I was You
[129:34] can You can take it away. Thanks, Dan.
[129:36] They managed to do what apparently I was
[129:38] the very first person to ever ask for.
[129:41] Um, the original exhaust, the main issue
[129:44] with it is that it a didn't match at all
[129:46] cuz it was like silver and like giant
[129:49] and ugly and b had a bunch of damage on
[129:51] it from uh either the previous owner or
[129:54] I both of us dropped the bike at some
[129:56] point or another. So, it had some damage
[129:57] from that. Um, so I needed a new exhaust
[130:00] and I was like, "Okay, so I want a quiet
[130:03] one." And they were like,
[130:06] "Nobody's ever asked for that before.
[130:09] What are you talking about? Nobody ever
[130:11] wanted to make their motorcycle quieter.
[130:14] And I was like, "Yeah, but but like I I
[130:16] I like I like my V twin. I like that it
[130:18] doesn't like scream. I like that it's,
[130:20] you know, modestly
[130:22] modestly loud." And um
[130:26] and anyway,
[130:27] >> really funny.
[130:28] >> Yeah. Yeah. So they they uh I found this
[130:31] carbon fiber pipe that was compatible
[130:33] with my bike and I was like, "Hey, what
[130:34] about this one? This doesn't like
[130:36] advertise being super loud." So they got
[130:38] it and they got a baffle for it. Um, so
[130:40] it has like a like it has like a deep
[130:42] growl, but it but it's not super loud. I
[130:45] actually really like it. It sounds
[130:47] amazing. Uh, John Josh in Canada asked,
[130:49] "What's your organ donor status? I'm a
[130:51] proud organ donor. If I don't need them,
[130:53] I'd be happy to have somebody else have
[130:55] them." So,
[130:58] yeah. I honestly I I had the I was I was
[131:02] riding riding scooters in Taiwan part of
[131:06] this last week and I genuinely had the
[131:09] thought of like would my o organ donor
[131:12] card work here? Um
[131:15] like what what happens if you die in a
[131:18] foreign country? Um
[131:21] I don't know. But I'm still alive so no
[131:23] real reason to keep thinking about it.
[131:25] Um,
[131:28] do we want to talk about the creepy dude
[131:29] stalking Rockstar's office?
[131:33] Sure.
[131:36] Uh, user then pomegranate-625,
[131:41] which sounds like the most Reddit
[131:42] generated autogenerated username ever,
[131:45] uh, posted on a G Oh, it's not even Oh,
[131:47] no. Subreddit. Yeah. Okay. posted on a
[131:49] GTA 6 fan subreddit this week about a
[131:52] decreased oxygen levels on the weekends,
[131:55] which this user believes is executives
[131:58] meeting after business hours at the
[132:00] headquarters to discuss marketing/tra 3.
[132:06] This same user has been tracking audio
[132:08] levels around the office to see if they
[132:10] can identify key topics and trends as
[132:13] well as having counted cigarette butts
[132:16] left outside to try to identify higher
[132:18] levels of stress. So, I'd like to jump
[132:21] in real quick here and say this is a
[132:23] prime example of how a behavior may be
[132:27] legal, but that doesn't mean that it
[132:29] demonstrates good social skills.
[132:32] >> Yeah.
[132:36] kind of weird.
[132:37] >> Oh my goodness.
[132:39] >> It's kind of It's funny.
[132:41] >> Yeah. I mean, it's funny until you're a
[132:43] rockstar employee who feels deeply
[132:45] uncomfortable about knowing that this
[132:47] guy is there and like
[132:50] >> you see him like you go outside for a
[132:52] smoke and then you see him like scurry
[132:54] up afterwards and he like picks up your
[132:55] cigarette with tweezers and gets it like
[132:57] >> Yeah.
[132:58] >> DNA scanned. like
[133:03] it's definitely wacky. Um yeah.
[133:06] >> Yeah. There's uh what what did we what
[133:08] did we used to what did we used to say
[133:10] back in the late 90s early 2000s? Get a
[133:13] life.
[133:14] >> Huh? Yeah.
[133:15] >> That that was before touch grass. That
[133:17] was that was what we used to say.
[133:19] >> Yeah.
[133:20] >> Get a [ __ ] life.
[133:22] >> Touch grass is pretty good.
[133:23] >> Touch grass is pretty good. But also no.
[133:25] Do more than touch grass. actually like
[133:28] >> grass and get a life.
[133:29] >> He might have touched grass in the
[133:30] process of this. He actually probably
[133:32] did.
[133:33] >> Yeah, that's crazy.
[133:35] >> So, it doesn't work anymore. Um, good
[133:38] news. Other topic, Ampear is back.
[133:40] Videoards.com with a zed uh reports that
[133:46] Yeah.
[133:46] >> has listed new GeForce RTX 3050 and 3060
[133:50] cards on their website. The 3060 uh
[133:53] M2521
[133:55] plus N630
[133:58] is the older 12 gig model with 3584
[134:02] CUDA corores at 1320. Uh
[134:05] >> I don't think you have to read every
[134:06] stupid
[134:08] >> 1777 MHz things you don't care about. Um
[134:13] there's also a 3050 coming. I think I
[134:16] mentioned that previously. Uh Manley's
[134:18] web page does not list North American
[134:20] distribution, but Tech Powerup reported
[134:23] uh that early May,
[134:27] but a Tech PowerUp report from early May
[134:29] names Asus, Colorful, and MSI as
[134:31] partners uh in at least the RTX 3060
[134:34] resurrection expected later this month.
[134:37] Uh the likely reason for the 3000 series
[134:39] returning versus the 4000 uh is foundry
[134:43] availability. The 3000 series GPUs are
[134:45] built on Samsung's 8nmter node, while
[134:47] the 4060 shares the same Nvidia 4nmter
[134:50] node as the 5060. So, yeah, that does
[134:53] make a lot of sense.
[134:54] >> I think that's a typo. I think they mean
[134:55] TSMC rather than Nvidia there. I'll fix
[134:57] that for you. So, just for context, the
[135:00] 3060 12 gig originally launched in 2021.
[135:08] >> Yeah, it's still a pretty capable card
[135:11] though. But like,
[135:12] >> but how sad is that that it's been 5
[135:15] years and that that
[135:17] >> is that video out
[135:18] >> and that the 3060 12 gig would be an
[135:21] acceptable gaming GPU today.
[135:25] No, I Yeah,
[135:26] >> like I remember when we used to get new
[135:28] GPUs every year instead of waiting five
[135:31] years for what did we get two
[135:33] generations in that time and with no
[135:35] 6000 series on the horizon. I don't
[135:38] think there's any rumors swirling which
[135:40] basically means we're 6 to 12 months out
[135:44] at least.
[135:47] There's a video coming from a I'm going
[135:50] to be incredibly vague here. A creator
[135:54] um who who like charted a bunch of Man,
[135:58] I'm going to have to just keep being
[135:59] really vague. I think it's not out yet.
[136:01] >> Um a bunch of things that have been
[136:03] happening over a bunch of years. Um
[136:07] >> Oh, no. Yeah, it's not that one. Yeah, I
[136:10] don't think it's out yet. Um but it's
[136:13] it's it's rough, man. since since
[136:18] you know things since it felt good.
[136:19] Basically, it's it's been bad. What a
[136:21] what a boring obvious statement, but
[136:23] it's like even worse than I thought.
[136:25] When you see it charted out,
[136:26] >> this sounds like a Paul video. I'm
[136:28] calling it.
[136:29] >> I I don't know. I can't say. Uh but but
[136:33] 3060 like I mean I've talked so many
[136:36] times on this show about uh somebody
[136:39] running Arc Raiders on a 1080 and then
[136:42] you get to a 30 and now you've got some
[136:44] you know you've got some rate tracing
[136:46] going on and some of the new software
[136:48] features and stuff a little bit
[136:50] >> little bit little bit of RA trace
[136:52] >> little bit
[136:53] >> some light DLSS
[136:55] >> little bit of things going on here there
[136:57] u 12 gigs of VRAM should also really
[136:59] help the 3050. I know,
[137:04] you know, people need affordable
[137:06] options, but the 3050 mostly because of
[137:08] the VRAMm really feels like it starts to
[137:11] become
[137:12] >> that's a pretty underpowered card today.
[137:14] >> Yeah. 3056 gig is like
[137:17] >> we're getting dangerously close to
[137:18] modern iGPUs being somewhat competitive.
[137:21] >> Totally.
[137:22] >> Yeah. So, that that's that's a lot less
[137:25] interesting to me. But that that 3060 12
[137:27] gig is is a is a serious card. you can
[137:30] get some stuff done with that. So,
[137:31] that's not too bad.
[137:34] >> Yeah,
[137:35] >> I'm I'm supportive of this overall.
[137:38] Anything that makes GPUs more plentiful
[137:41] and or more affordable, I am I am all
[137:44] for. I wish this wasn't the solution cuz
[137:47] this is effectively Nvidia scrapyard
[137:50] wars.
[137:51] Like, we're we're manufacturing old GPUs
[137:55] instead of going on Facebook Marketplace
[137:57] to obtain them. But fine,
[138:00] >> the fabled uh just buy everything from
[138:02] the store scrapyard worse.
[138:04] >> Yeah,
[138:05] >> if this is what it takes to get slightly
[138:08] more affordable GPUs, then I guess I'll
[138:10] take it.
[138:12] Well, I think one thing that I'm really
[138:14] hoping to come from this is just that
[138:15] people that just need a GPU that say has
[138:19] 12 gigs of VRAM and can do a few things,
[138:22] uh, there will be more options for that
[138:24] because if there's more total output of
[138:27] graphics cards, people that need at
[138:29] least that amount of performance will
[138:30] have more options. Hopefully, that'll
[138:32] ease up the used market, do some things,
[138:34] help prices here and there.
[138:36] >> Will any of that actually happen? kind
[138:38] of doubt it, but I can hope.
[138:41] >> Hey, here's a fun one. Nintendo of
[138:43] Europe to pay €35 million in fines due
[138:47] to Joy-Con drift. In response to faulty
[138:50] JoyCons, a complaint from a French
[138:51] consumer organization was filed claiming
[138:53] that Nintendo of Europe misled customers
[138:55] by failing to disclose Joy-Con
[138:57] controller defects transparently, which
[138:59] caused consumers to buy unnecessary
[139:01] replacement controllers. Nintendo has
[139:03] agreed to pay $35 million, about $40
[139:06] million US, after a French regulator
[139:08] found them liable. Nintendo of Europe
[139:10] said in a statement that it did not
[139:12] intentionally mislead customers and that
[139:14] this settlement does not constitute an
[139:16] admission of guilt and reflects only the
[139:18] amicable resolution of legal
[139:20] proceedings.
[139:22] According to a UK watchdog, over 40% of
[139:24] Switch Joycons have experienced stick
[139:27] drift. It is remarkable to me that a
[139:30] company of the size of Nintendo managed
[139:32] to not fix that.
[139:36] >> Yeah.
[139:36] >> I I I don't
[139:38] >> pretty wild.
[139:39] >> I don't get it. Um
[139:43] and like that my own experience with
[139:45] JoyCons backs this up. Like I think over
[139:49] 40% of our first generation JoyCons
[139:51] drift now. And it's not like they've
[139:54] been used super extensively. like
[139:57] they've been they've been used, but I
[139:59] think
[140:01] three of us have like finished Breath of
[140:03] the Wild.
[140:06] My son played one of the Pokemon games
[140:08] on Switch. Like, like there's there's
[140:11] maybe if I were to add it up like 400
[140:14] hours of gaming across three sets of
[140:17] JoyCons. Like, it's not a ton.
[140:21] >> Pretty rough.
[140:24] Uh, a unique username says 60% of people
[140:27] have a life outside of playing Nintendo
[140:28] games. Uh, no, because you don't even
[140:30] you don't even have to play that much
[140:32] for them to Oh, for them to break. Oh,
[140:34] yeah. I guess. So, they have a life that
[140:36] basically does not include Nintendo
[140:38] games. Cuz it's a fun from my
[140:40] understanding, it's like a fundamentally
[140:42] flawed design, right? Like it will
[140:44] drift. Is that correct? It's just a
[140:47] matter of when it dies.
[140:50] >> I think so. L explains joke makes it not
[140:53] funny.
[140:55] >> Sorry.
[141:02] >> Sorry. I got to read something for a
[141:03] sec. Um, what do we have next though? Do
[141:06] you want to do Apple introduces Siri AI?
[141:09] >> Yeah, I can do that.
[141:11] At WWDC on Monday, Apple announced its
[141:14] second attempt at Apple Intelligence,
[141:17] which it is calling Siri AI, but it's
[141:20] Gemini under the hood. According to
[141:22] Apple, the new Siri AI can draw on
[141:24] personal context from your messages,
[141:26] emails, and photos. Get up-to-date
[141:27] answers from the web. Use on-screen
[141:29] awareness to tailor its assistance based
[141:31] on the app you're using. Use all of that
[141:33] information along with apps on your
[141:35] device to perform tasks on your behalf.
[141:37] Craig Federigis took a pointed shot at
[141:40] other AI companies during the keynote,
[141:41] saying, "Some are racing forward,
[141:44] seemingly pursuing AI for the sake of AI
[141:46] without clear regard for the people it's
[141:48] meant to serve." But according to Craig,
[141:50] Apple's approach is to focus on making
[141:51] AI that improves the lives of their
[141:54] customers. Apple also attempted to
[141:56] differentiate themselves from other
[141:57] companies by saying that everything is
[141:59] built privacy first with anything that
[142:02] leaves the device handled by private
[142:03] cloud compute which the company says
[142:05] does not store or share user data and
[142:08] can be audited by outside experts. A
[142:11] presenter demoed Siri AI by asking it
[142:13] about a concert in San Francisco. Sorry,
[142:15] I'm going to short aside here. What is
[142:17] it with every big tech company using
[142:21] going to concerts as a demo lately?
[142:23] Google just did it. Okay, just Apple and
[142:26] Google. Fine, whatever. But it feels
[142:27] like it comes up a lot. Are concerts not
[142:31] like really really expensive now?
[142:36] Yeah,
[142:37] >> this feels like you know how you know
[142:40] how back in the day Apple would would
[142:42] show you all their new calendar features
[142:44] during the keynote and like the example
[142:47] calendar they have has like wake up and
[142:51] do yoga and then just like a bunch of
[142:53] chill. They have like four events in a
[142:56] given day and I'm like, "Yeah, this is
[142:58] why I can't use your calendar because my
[143:00] calendar will have like 25 things on it
[143:02] and your your data density is so low on
[143:04] your calendar that I like can't see any
[143:06] of it at once." Like it does it not does
[143:09] it just feel like the executives are
[143:12] just building the products to do the
[143:14] things that are relevant to them and to
[143:16] no one else? Like as Apple's exa as
[143:20] Apple's as exa as Apple's seuite has
[143:23] aged their focus on like health related
[143:26] products and fall monitoring stuff seems
[143:28] to have aged a seems to have advanced
[143:30] along with them. Like I I don't know.
[143:32] It's just kind of a funny observation.
[143:34] You know what? I'll move on. Um,
[143:37] anyway, they asked it to about a concert
[143:39] in San Francisco, asked it to set a
[143:42] reminder to sign up for the ticket
[143:43] lottery, and then pull a friend's new
[143:45] address out of a message and generate
[143:46] driving directions. Pretty cool, pretty
[143:48] useful. A second demo had Siri plan a
[143:51] World Cup watch party for the Brazil
[143:52] versus Morocco match, getting Siri AI to
[143:55] plan a menu for the event, including a
[143:56] dessert idea, which was pulled out of a
[143:58] text. Some other things were demoed uh
[144:00] including improved visual intelligence
[144:03] capable of doing stuff like splitting a
[144:04] bill with Apple Cache or surfacing
[144:06] nutritional information from a plate of
[144:08] food. Also, better dictation and more
[144:10] customizable voices. Apple is also
[144:13] launching a dedicated Siri app that uses
[144:14] iCloud to privately sync your
[144:16] conversation history across your
[144:17] devices. Right now, Siri AI is available
[144:20] for developer testing and will launch as
[144:21] a public beta later this year.
[144:27] Discussion question. Do you think this
[144:29] approach to AI is better than other
[144:31] companies?
[144:35] >> H
[144:40] uh well I was reading other stuff during
[144:41] a decent amount of this. It looks like
[144:43] it's local first and and privacy focused
[144:46] and stuff and that seems
[144:47] >> beneficial. Good.
[144:49] >> It's funny. I uh I I took a lot of flack
[144:51] when I was talking about the uh RTX
[144:53] Spark launch in our in our video walking
[144:56] in the park in in at Comp Computex like
[144:59] uh you know Lionus is changing his tune.
[145:01] It's like well yeah when I'm presented
[145:03] with new information I I change my
[145:06] perspective. Um if I didn't do that you
[145:08] should not listen to me at all. Um I I
[145:12] like local AI. Local AI seems pretty
[145:18] cool. I don't like, you know, Sam Alman
[145:21] listen Sam Alman listening in on
[145:23] everything that I'm, you know, confiding
[145:26] in my AI. Um, that I'm super not into.
[145:29] But anything that I can run locally, I'm
[145:31] actually,
[145:33] as long as it's legitimately useful, I'm
[145:36] Yeah, I'm kind of excited about. There's
[145:38] a lot of There's a lot of really cool
[145:40] stuff. Nvidia's demos were really cool.
[145:42] Apple's demos here look pretty
[145:44] compelling. I need to see it actually
[145:46] work because I've I've heard all of this
[145:48] before and the like the jaded part of me
[145:52] is like, "Okay, what's going to be
[145:53] different this time?"
[145:56] >> But if it's cool, then it's cool.
[145:59] >> The demons through the portal, it's like
[146:02] um it's already here. I I I think like
[146:05] local if if you hate it, local is the
[146:08] way to go. If if that makes sense. Um
[146:12] because you're, you know, don't don't
[146:15] pay for anything at that point other
[146:17] than hardware. I guess hopefully you can
[146:18] use the same hardware you already have,
[146:20] but if you're not fortunate to have
[146:21] super expensive computer hardware, then
[146:23] I guess you need to buy computer
[146:25] hardware. You can try to get off the
[146:26] used market or something to to do
[146:28] whatever. But anyways, um by by going
[146:31] the local route, you can keep up to
[146:34] speed
[146:35] um but not directly be paying
[146:41] uh or or feeding data into some of these
[146:44] massive uh AI companies.
[146:48] So, if if that's your biggest gripe,
[146:50] which I think it's the biggest gripe for
[146:52] a lot of people, um you can you can
[146:53] avoid that with local. And with Odysius
[146:56] being out there, Local is more
[146:58] approachable than ever. Um and as far as
[147:00] I can tell, it's just going to keep
[147:02] being more and more approachable over
[147:05] time, which is which is fantastic.
[147:06] >> I can relate to a lot of the other
[147:08] gripes. I mean, the fact that the fact
[147:11] that this whole technology was built
[147:13] with data that was stolen from
[147:16] >> Yeah. creators, including me, by the
[147:19] way, like me personally.
[147:20] >> I
[147:22] >> the fact that I was able here. Okay,
[147:24] this
[147:26] >> probably could use.
[147:27] >> I mean, it's it's also true for like
[147:29] anyone who had a Reddit account. Like
[147:30] this isn't this isn't necessarily like
[147:32] special to any individual person. Like
[147:34] if you were a member of the internet,
[147:36] your information was stolen. Like this
[147:38] could probably use a little bit more
[147:39] context, but I managed to get OpenAI to
[147:42] generate this picture before the show
[147:44] with basically no difficulty whatsoever.
[147:50] >> Sorry, I just got jump scared by Sandy.
[147:53] How long have you been there?
[147:57] Holy crap.
[147:59] What the heck?
[148:01] >> That was like some poltergeist stuff
[148:03] right there. That was amazing.
[148:05] I had no idea he was there. Oh my god.
[148:10] Wow. Thanks, Sammy. Sorry, lights.
[148:13] >> No, you're good.
[148:16] >> Whoa.
[148:17] >> Um, and that image that I generated was
[148:19] in spite of chat GPT refusing multiple
[148:22] times to generate it on the basis of
[148:25] that it was against its
[148:28] >> uh what what did it say? Um yeah,
[148:31] against its its terms. Uh, so it was
[148:34] clearly clearly trained on like like me
[148:38] like my actual creative output. Um, and
[148:41] and probably almost every single one of
[148:44] you as well. And that's that's crap.
[148:48] That's pretty crap.
[148:49] >> No, I and I I completely agree with
[148:50] that. My
[148:51] >> But it's also out there.
[148:53] >> Yes. My thing is that that happened and
[148:56] that can be terrible and we can all
[148:58] agree that that's terrible and we can
[149:00] all agree that these big companies are
[149:02] profiting off of looting all of that
[149:04] from all of us. Um,
[149:08] and I think the biggest way if you just
[149:10] really want to get back at them for it,
[149:12] the best way to do that is local. run it
[149:16] in a way that they profit zero and gain
[149:20] data zero
[149:22] >> because then at least
[149:23] >> you might as well use the stuff that was
[149:25] stolen from you.
[149:27] >> Yes, it's using things that was stolen
[149:28] from other people as well and that's
[149:29] that's its own ethical
[149:31] >> and that sucks
[149:32] >> quandry
[149:33] >> and that's pretty crap.
[149:38] >> I don't have much more to say on that to
[149:40] be completely honest. Brutal. But um
[149:46] I I think people are very actively I
[149:49] know quite a few people right now that
[149:51] are using this stuff to accelerate
[149:52] themselves in ways that is making them
[149:55] much more competitible sorry much more
[149:57] competitive than their peers
[150:00] competitive.
[150:02] Um, yeah. Uh, and if you can do that
[150:07] with local systems, then you're able to
[150:09] keep up with those people and not profit
[150:12] these these big organizations. Um,
[150:15] and I I think that's a cool at least
[150:18] option. I at least option at the very
[150:21] least I think you can go with the like
[150:23] the performant combo of having the um
[150:29] Oh, what are why why did I blank on the
[150:31] name? the Frontier models um and using
[150:36] that for like special high difficulty
[150:39] stuff if you want and then having the
[150:40] local models for for everything else and
[150:43] you can do some really I hear people
[150:45] talk about how the Frontier models are
[150:47] so much further ahead than the local
[150:48] models. I think that's overplayed a
[150:50] little bit. the frontier models are are
[150:52] quite clearly ahead. But I think in the
[150:55] use cases that as far as I can tell most
[151:00] people are using AI for you can really
[151:02] do a lot of that with local models.
[151:05] >> Um and then as I said earlier in the
[151:08] show when you really do need to push it
[151:09] you can you can
[151:11] >> switch your your uh model to one of the
[151:14] frontier models. You can do that in in
[151:16] Odysius. can have both cloud-based
[151:18] frontier models, whatever, and you can
[151:21] have local models running through the
[151:23] same UI. Um, so yeah, own it yourself. I
[151:26] think that's the only real way. Um,
[151:32] yeah,
[151:34] >> one day after a group of people looted
[151:36] the library of Alexandria, they took it
[151:38] all, but somehow you're able to get a
[151:41] copy of that. I think you should take
[151:43] it.
[151:48] Ah, no, it's not really related. Um, one
[151:50] day after its discovery, Meta has pulled
[151:54] the unreleased facial recognition system
[151:57] out of its smart glasses companion app.
[152:00] Um, Wired revealed that the code was
[152:02] sitting on more than 50 million phones.
[152:05] The system was internally called Name
[152:08] Tag and according to Wired, it was
[152:10] designed to convert faces captured by
[152:12] the glasses into unique biometric
[152:14] signatures. then compare those face
[152:16] prints against a database stored on the
[152:19] user's device. Wired also reported that
[152:22] any faces the system failed to recognize
[152:24] were cropped, indexed, and stored
[152:26] locally on the phone for future
[152:28] processing. Note that Meta refused to
[152:32] answer Wired's questions on whether the
[152:34] biometric data would be sent back to
[152:36] Meta's servers. The feature was never
[152:39] turned on, but the machinery was already
[152:41] built into the Meta AI app as early as
[152:43] January, even as Meta publicly said it
[152:46] had made no final decision about face
[152:49] recognition. name tag first surfaced in
[152:52] the news in February when the New York
[152:53] Times cited internal meta documents
[152:56] showing the company was weighing a
[152:58] launch as soon as this year with one
[153:00] memo reportedly describing releasing it
[153:02] during a dynamic political environment
[153:04] when privacy and civil liberties
[153:06] advocates would be distracted.
[153:11] >> Yeah.
[153:13] After Wired's first report, MetaVP of
[153:16] communications Andy Stone said the
[153:17] company could not answer questions about
[153:19] how the system would work because, this
[153:21] is a quote, "The feature does not
[153:22] exist."
[153:25] The updated Meta AI app strips out the
[153:27] face recognition software.
[153:30] You doing okay over there, Dan?
[153:33] >> This current year is absurd.
[153:36] >> This is just
[153:40] Do you need You know what? Remove him.
[153:42] Remove the thing. Okay,
[153:43] >> remove the thing. It counts. It counts
[153:45] if Dan If Dan can't anymore. Um, the
[153:48] updated meta app strips out the face
[153:50] recognition software, including the code
[153:52] that ran the name tag recognition
[153:53] process, the person recognized alert
[153:56] that the app would have shown, and the
[153:57] folder where the app would have stored
[153:59] cropped images and biometric signatures
[154:00] of unidentified faces. Our discussion
[154:03] question is, is there anything we can do
[154:05] to protect our privacy when this kind of
[154:07] tech simply even exists? Um,
[154:12] I mean, there's those uh there's those
[154:15] cool camera rejecting reflective
[154:18] accessories you can wear. I don't think
[154:21] it helps if they aren't emitting IR. So,
[154:24] uh, trying to think what else we got.
[154:28] Not go in public, says Scrappy DP. Way
[154:31] ahead of you there. If I wasn't a
[154:33] YouTuber, I would just hang out in my
[154:34] house.
[154:36] >> What else we got?
[154:37] >> If you're in semi-private places,
[154:40] Um, I don't know why I'm saying
[154:42] semi-private. I mean private places, but
[154:44] I'm not necessarily including only like
[154:46] your house. If you're in like say your
[154:48] your friend's house, right?
[154:50] >> Like that or or a business like a a
[154:53] restaurant.
[154:54] >> Um, you might be able to make it the the
[154:56] rule of the the the business or
[154:58] restaurant or those are the same thing,
[155:00] whatever. Uh, or maybe a house rule um
[155:03] that you can't wear these glasses when
[155:06] you're in those places. The the tough
[155:08] part is this rhymes to me with the
[155:12] palunteer saving a bunch of lives and
[155:14] also stealing all of your precious data
[155:16] that you could ever have uh in in
[155:18] return, which is it's going to be really
[155:22] hard to tell people, hey, you can't wear
[155:24] your glasses that you might need to see.
[155:26] >> Yeah. Cuz they'll they'll have a
[155:28] prescription in there and they might not
[155:30] have prescription not smart lenses
[155:32] because these things are you can you can
[155:34] hate it all you want, but it's going to
[155:35] go mainstream. like it's going to go
[155:38] people just have
[155:38] >> Yeah. They might not have their other
[155:40] glasses on them. So, they might just not
[155:42] be able to see if you want your privacy.
[155:46] >> Yeah.
[155:46] >> Uh which is these these cost exchanges
[155:49] are going to make it very socially
[155:50] difficult. It would be very socially
[155:52] difficult to say no to the palunteer
[155:54] thing in the in the eyes of of someone
[155:57] potentially dying. It would be very
[155:58] socially difficult to say, "No, you
[156:00] can't wear your glasses." to someone who
[156:02] actually needs them to be able to see
[156:03] and oh, they just also happen to be able
[156:05] to surveil everyone that's in your
[156:07] location. Uh, the other problem is you
[156:11] can definitely just mod them so that the
[156:13] indicator light does not go on when you
[156:14] record video. So, like it's over
[156:18] already. It barely started and it's
[156:20] over. If someone has these glasses, you
[156:23] have no real way of knowing uh they
[156:25] might be recording you. like it just
[156:28] it's done. That includes in bathrooms,
[156:30] that includes in um a huge variety of
[156:35] places. I I'm trying to imagine like
[156:38] um there are a variety of places around
[156:41] the world. I was there recently, so I
[156:42] was just thinking of it recently. Um but
[156:46] even here in Taiwan, they have certain
[156:48] hot springs that are like nude hot
[156:49] springs. In in Japan, they have the the
[156:51] onsons where you're not really supposed
[156:53] to wear clothes. Mhm.
[156:54] >> Um I'm assuming you would still be able
[156:56] to wear glasses.
[156:59] >> What if you can't see without your
[157:00] glasses? You would have to be able to
[157:01] wear glasses.
[157:02] >> Yeah.
[157:04] >> Um
[157:07] I suspect cameras aren't exactly welcome
[157:10] at like nude beaches. There's nude
[157:12] beaches in North America. I'm not just
[157:14] trying to centralize this to Asia. Um
[157:17] but someone would be allowed to wear
[157:18] their glasses. The thing is like glasses
[157:21] can get in a lot of places that cameras
[157:23] can't. Um,
[157:26] and are much more socially acceptable to
[157:28] point at things because you're pointing
[157:30] at it with your face. Um, oh man. MCH
[157:34] says, uh, Luke and Lionus, time to trim
[157:37] the people you hang out with. It's like
[157:40] I I wish it was that simple, you know?
[157:43] Like I, um,
[157:46] it's not that simple. Like with
[157:50] acquaintances,
[157:52] sure, maybe you could do that. But with
[157:55] like longtime friends, are you're you're
[157:59] going to terminate a friendship over
[158:00] someone wearing AI glasses? Some people
[158:03] might. Most people won't. What about
[158:06] family? Most people won't. Are you going
[158:07] to excommunicate family over them
[158:10] wearing AI glasses?
[158:11] >> This is the exact thing I just mentioned
[158:13] where uh it's it's socially difficult.
[158:16] That's what I meant by this. Yeah,
[158:18] >> it it is going to be socially difficult
[158:20] to the point of not possible for a lot
[158:22] of people um to
[158:26] do that.
[158:27] >> Mind paradox uh flagged the solution for
[158:30] us here. Um
[158:33] great moment from the Princess Bride. I
[158:35] don't know if you guys have audio.
[158:39] >> Oh lordy, why do I suck at this?
[158:45] The the Yeah, fixed. There you go. There
[158:48] we go. Ready?
[158:50] >> Why you wear a mask? Will you burn the
[158:52] acid or something like that?
[158:54] >> Oh, no. It's just they're terribly
[158:55] comfortable. I think everyone will be
[158:57] wearing them in the future.
[158:59] >> It's a great line. Masks are terribly
[159:02] comfortable. I think everyone will be
[159:03] wearing them in the future. So, because
[159:05] he asks why he wears a mask. Um, yeah.
[159:09] So, that's an option. You could just
[159:10] everyone could just wear masks.
[159:16] Also, high social friction. I think when
[159:18] you walk into the bank in your bandit
[159:20] mask, they probably will ask you to
[159:21] remove it.
[159:27] Yeah. Grwing says, "I can't even get
[159:29] family to care about basic web privacy.
[159:31] Little chance I can get them to care
[159:32] about the privacy implications of
[159:34] devices like this given an interesting
[159:35] feature set." Yeah, exactly.
[159:40] >> Um,
[159:42] yeah. Oh man, I didn't even think about
[159:44] like co-workers. I sis says, "The CEO at
[159:47] my old company wore the Facebook Oakley
[159:49] glasses to the office regularly." Yeah.
[159:51] So, friends, family. Yeah. Don't forget
[159:53] about like your boss. Well, you're going
[159:54] to you're going to tell your boss,
[159:56] "Okay, you wear you wear Facebook
[159:58] glasses. I quit."
[159:59] >> Some people might have the luxury to do
[160:01] that. I guarantee you not everybody
[160:03] does.
[160:04] >> Most people won't. The vast majority of
[160:06] most people won't. Um, man, somebody
[160:09] said something in here. Where'd it go?
[160:13] No.
[160:18] Ah, I someone said places are
[160:20] successfully banning them. Um, and I I
[160:23] think I mentioned earlier that like you
[160:25] might be able to ban them in certain
[160:26] businesses and private places as well,
[160:29] >> but I think Lionus' for now is really
[160:32] important here. I'm not necessarily
[160:34] certain why he's saying for now, but my
[160:36] reason for for now. Um, is
[160:41] what about when it's not just meta?
[160:44] >> How do you keep up with all of the
[160:46] different glasses that are smart
[160:47] enabled?
[160:48] >> Dude, look at how many look at how many
[160:50] like tablets there are on AliExpress.
[160:54] How many brands of smart watches there
[160:57] are. You give it another five years and
[160:59] there's going to be literally a thousand
[161:03] different options for smart glasses.
[161:04] They are going to be smaller and sleeker
[161:07] than anything that we're looking at
[161:08] today.
[161:10] >> It's going to become very difficult to
[161:11] tell if there's a camera in it at all.
[161:13] Um,
[161:14] >> and in parallel,
[161:17] people are going to
[161:20] just sigh and shrug and go, "Well, it's
[161:22] inevitable." and are going to be less
[161:26] vigilant and it's just going to kind of
[161:29] they're going to get harder to notice
[161:31] and people are going to care less and
[161:33] those points are going to intersect and
[161:35] they will just kind of be everywhere and
[161:36] I hate to be defeist about it and by all
[161:39] means you should you should advocate
[161:42] strongly in your social circle for
[161:44] people you know not wearing them in
[161:46] situations where there's an expectation
[161:48] of privacy um and advocating for
[161:51] yourself and what you're comfortable
[161:53] with I'm I just I won't be super
[161:56] surprised when it maybe doesn't work.
[162:01] >> Um what will work is telling you about
[162:03] our next batch of sponsors, though.
[162:06] >> The show is brought to you by
[162:06] Squarespace.
[162:08] >> Your website says a lot about who you
[162:11] are, how you operate, and the types of
[162:13] customers you wish to attract. So, it
[162:15] shouldn't look like this. It should be
[162:16] easy to understand and navigate.
[162:19] Thankfully, our sponsor Squarespace
[162:21] makes it easy to build a website for
[162:22] your business that does just that. To
[162:26] get started, you just pick from one of
[162:27] their stylish templates or create
[162:29] something more custom with their design
[162:31] intelligence tool. It's more than just
[162:33] looks, though. There's plenty of tools
[162:35] under Squarespace's hood that will help
[162:36] with things like SEO and monetization.
[162:39] From there, you can keep track of how
[162:41] your business is doing, adjust your
[162:42] website on the fly, and make sure that
[162:44] potential visitors get the most out of
[162:46] what you have to offer. We've even used
[162:48] Squarespace for our Linus Media Group
[162:50] website for many, many, many years now.
[162:53] So, start building your website today
[162:54] and get 10% off your first purchase by
[162:57] visiting squarespace.com/wan.
[163:01] Our last sponsor for today is actually a
[163:03] site that I was on yesterday because I
[163:06] was playing around with my Naz. Dan, you
[163:08] can show that picture.
[163:13] Luke, this is why my GPU wasn't showing
[163:17] up.
[163:19] I completely forgot that when we built
[163:21] it, we planned to open it back up at a
[163:24] later date and add like an Oculink to
[163:28] PCIe or M.2 to PCIe riser or something
[163:32] to run this GPU as a compute device and
[163:35] then just
[163:37] never got to it.
[163:39] So, it was literally wasn't even plugged
[163:41] in.
[163:43] Well, there's your problem.
[163:44] >> Anyway, I fixed that yesterday. Uh, and
[163:47] I was firing up server parts deals
[163:48] because I was like, "Oh, yeah, like
[163:50] what's um how's my how's my server
[163:53] capacity? Do I need any more drives?"
[163:54] And they're a great place. Okay, sorry.
[163:56] I'll read the thing. Uh, wait lists for
[163:58] the newest hard drives can sometimes be
[164:00] months long. And if you're not a
[164:01] hyperscaler, you're probably not a
[164:03] priority. Well, Server Parts Deals has
[164:06] shipped tens they have tens of thousands
[164:08] of drives in stock and even orders of a
[164:11] 100 drives or more can often be shipped
[164:14] same day. Everything is tested once they
[164:17] get them, even reertified Western
[164:19] Digital and Seagate drives. They get an
[164:21] initial smart test, FIO throughput
[164:23] across multiple parts of the disc, a
[164:24] full surface badlock scan, and then
[164:27] another smart to make sure that nothing
[164:28] changed. For larger businesses or data
[164:30] center buildouts, they also do buybacks.
[164:32] or if you're upgrading your system, they
[164:34] can help offset the cost. Do you need
[164:36] hard drives fast? Skip the wait list
[164:38] with server parts deals. We'll have a
[164:40] link down below. Sorry, it's server part
[164:44] deals, isn't it? Hold on, I'm going to
[164:45] double check. I can never remember.
[164:47] >> I thought it was parts, but my talking
[164:49] points say part.
[164:52] >> It's part.
[164:53] >> It is part. It is part. Don't go to
[164:55] server parts deals. I don't know what
[164:56] the heck that is. Server part deals.
[165:00] >> Good job, Lionus. Good podcast sponsor,
[165:02] Reed. Way to go. Screw up the name of
[165:04] the company.
[165:06] Thank you for that, Dan. That's very
[165:08] helpful.
[165:08] >> He's a professional. He's been doing
[165:10] this for 12 years.
[165:12] >> All right. Uh, okay. You got to do the
[165:13] next topic cuz I got to run to the
[165:14] washroom.
[165:16] >> Okay. Sure. Uh, one way to get real
[165:18] confused about pricing is to load up
[165:20] server part deals and have it
[165:21] automatically have the currency be in
[165:23] New Taiwanese dollars and have a hard
[165:25] drive cost $20,000.
[165:28] Uh, anyways, I was like, whoa, what
[165:32] happened?
[165:34] We're good. It's okay. Um, yeah, sure.
[165:38] We'll do this one. Uh, no. Los probably
[165:40] want to talk about that. Let's see.
[165:42] Let's see. Let's see. Let's see.
[165:47] >> Or we could get a couple comms out of
[165:49] the way.
[165:49] >> Yeah. This one's interesting. Sorry.
[165:50] >> Or we could get a couple comms out of
[165:52] the way for you specifically. Or do a
[165:53] topic, whatever you want. I don't think
[165:55] it'll be
[165:55] >> If you've got comms for me, we can do
[165:57] that. Yeah, sure.
[165:58] >> Yeah, why not? Let's see.
[165:59] >> If not, I got a topic here.
[166:00] >> Okay, cool. Um, if you haven't heard,
[166:03] there's a major Arch AUR supply chain
[166:05] attack happening right now.
[166:08] >> I have heard about that. I don't know
[166:10] all the details. Um, but yeah, I have
[166:13] heard about that. That sucks. Um, I
[166:16] think
[166:19] I feel like the server side of Linux is
[166:24] probably pretty well monitor monitored
[166:27] in terms of security, but with the user
[166:30] base on the desktop side being as small
[166:32] as it has been for years, I I don't
[166:35] know. Um
[166:37] it's it's going to get stress tested a
[166:40] little bit as there are waves and waves
[166:43] of people moving from from Windows over
[166:46] there and the the honeypot of being able
[166:49] to attack uh desktop Linux systems
[166:54] grows. Um
[166:57] it it's going to become more and more
[166:59] attractive for attackers. So we'll see
[167:03] how it all goes. Uh I I had at least one
[167:06] person reach out to me being like, "Are
[167:07] you okay cuz you're on Cashy?" I guess
[167:10] somewhat luckily for me. Um
[167:15] my my laptop is not my laptop is running
[167:18] Mint and Windows. My desktop is running
[167:20] Cache and Windows and I've been
[167:23] traveling for the last two and a half
[167:25] weeks. So my desktop is at home and
[167:27] powered down. So it has it has not hit
[167:30] me and as far as I've heard it's been
[167:32] fixed since then. So yeah, uh someone in
[167:35] chat said basically someone was uh
[167:37] sorry, S. John Trombley in chat said
[167:39] basically someone was adopting orphaned
[167:42] AUR packages and adding malicious calls
[167:45] to NPX
[167:47] uh in the install scripts.
[167:53] Yeah.
[167:56] Yeah. I
[167:59] I know a lot of people love the AUR
[168:03] I tend to try to lean
[168:07] away from it. I don't even know why.
[168:09] I've never actually had a reason. This
[168:10] was not a reason. And honestly, moving
[168:12] forward, I don't think this would be a
[168:14] reason. Um,
[168:17] I've just liked other ways of installing
[168:19] things so far, but you know, you do what
[168:22] you got to do sometimes. Anyways, you
[168:23] have another one?
[168:25] >> Yeah. What's your LLM of choice?
[168:28] >> Do
[168:30] flavor. uh flavor of the month.
[168:33] >> Yeah, kind of.
[168:35] >> Uh and if you're running like uh if
[168:38] you're running like T3 chat, it's really
[168:40] easy to skip through and use whichever
[168:43] one you want. Um they have a huge
[168:45] variety of options and they keep them up
[168:47] to date. Um I I find that's kind of a a
[168:51] fun way to use things is is T3 Chat
[168:53] because you can try so many different
[168:56] things out. Um and they have so many
[168:58] different options. Um, if you want to
[169:02] stay on top of things, that can be good.
[169:04] If you're not running local stuff, um,
[169:06] if you're running local stuff, if you're
[169:08] running something like Odysius, uh, you
[169:10] can you can cookbook every once in a
[169:12] while and see what things are good for
[169:14] you to run right now. I I don't
[169:16] necessarily even recommend having like
[169:19] one specific model of choice. Um, I
[169:22] think you kind of want to keep keep
[169:25] pushing, keep keep progressing with the
[169:28] with all the things. If you're if you're
[169:30] like up with the times enough to be
[169:33] using AI stuff in general, you might as
[169:35] well be up with the times enough to be
[169:38] um mobile with models to make sure that
[169:40] you're using the right stuff, the best
[169:41] stuff you can.
[169:44] >> Yeah.
[169:46] Do we want to talk about Canada's under
[169:48] 16 social media ban?
[169:52] >> Sure. I don't know a ton about it, but
[169:54] yeah. The Canadian federal government
[169:55] has introduced Bill C34, legislation
[169:58] that would ban kids under 16 from using
[170:01] social media platforms, particularly
[170:02] Facebook, Instagram, and Snapchat. The
[170:05] bill defines social media broadly as any
[170:07] site or app that lets users access and
[170:09] share content across provincial and
[170:11] global borders. Companies can get a pass
[170:14] if they set up proper safety measures
[170:16] for kids. However, the exact criteria
[170:18] will come later via official regulation.
[170:21] If passed, it would also put the
[170:22] responsibility on websites to protect
[170:25] children from harmful content, including
[170:27] cyber bullying. The bill would also
[170:30] create a digital safety commission to
[170:32] monitor sites for compliance, conduct
[170:34] audits, issue compliance orders, and
[170:35] dish out fines of either 3% of a
[170:38] company's global revenue or 10 million
[170:41] CAD, whichever's higher. Interestingly,
[170:44] AI chat bots won't be banned for kids
[170:47] under 16, but companies will need
[170:49] measures in place to deal with users
[170:50] expressing intent to commit violence or
[170:52] self harm. The bill would not, however,
[170:54] require chatbot companies to report
[170:56] these interactions to police.
[171:00] So, our discussion question is, unlike
[171:02] Australia's flat ban, this bill, if
[171:05] passed, would let platforms earn an
[171:07] exemption by proving adequate safeguards
[171:09] that don't exist yet. Is this a
[171:11] genuinely smarter way to incentivize big
[171:13] platforms to fix infinite scroll and
[171:15] addictive algorithms? Or is it just a
[171:16] lobbying backdoor where the exemption
[171:18] criteria gets watered down until
[171:20] everyone qualifies and nothing changes?
[171:23] I think only the time will tell. Uh,
[171:26] Electrobot asks a really good question.
[171:28] Maybe this is a more interesting
[171:29] discussion question. Why is this the
[171:30] only time we've ever heard of a fine
[171:32] that is a percentage of global revenue?
[171:34] That seems pretty smart. Um,
[171:36] >> that does seem pretty smart.
[171:37] >> If you want companies to actually, you
[171:39] know, stop doing a thing.
[171:41] >> Yeah. And the whichever is higher is
[171:43] like
[171:46] usually it's like whichever limit you
[171:47] reach first the opposite. Um, it's kind
[171:51] of fun. I have,
[171:54] it feels even weird to admit,
[171:57] but I have some hope in this being a
[172:02] genuinely smarter way to incentivize big
[172:04] platforms to fix infinite scrolling
[172:06] addictive algorithms over the loophole
[172:08] thing. The current Canadian admin has
[172:10] been doing, you know, not every step is
[172:13] is perfect, but has been doing some good
[172:16] stuff. And
[172:19] I have I have hope for the first time in
[172:21] a long time. So,
[172:22] >> are you ready for the other shoe to
[172:24] drop?
[172:25] >> Oh boy.
[172:26] >> Signal Duck.Go among firms weighing
[172:30] Canada exit over lawful access bill.
[172:33] >> Oh yeah. Yeah, good point. Yeah, this
[172:34] one blows. The big names in the VPN and
[172:37] privacy industry are threatening to pull
[172:39] out of Canada over Bill C-22, the lawful
[172:43] access bill, which would force service
[172:45] providers to retain user metadata for up
[172:47] to a year and build in ways for police
[172:50] and cis to access it. While the bill
[172:53] mainly targets core providers like the
[172:56] big satellite and telecom companies,
[172:58] VPNs are involved as the public safety
[173:01] minister can issue secret ministerial
[173:04] orders forcing any provider to comply.
[173:07] These orders only need sign off from the
[173:10] intelligence commissioner. No judicial
[173:12] warrant needed and providers aren't
[173:14] allowed to publicly say that the order
[173:16] exists. Signal told parliament that it
[173:19] would rather leave than betray its
[173:21] users. Duck.Go confirmed that it would
[173:24] pull its VPN, and NordVPN said it would
[173:26] consider all options, including leaving
[173:30] Canadian jurisdiction. Canadian-based
[173:33] Windscribe was the bluntest, saying it
[173:36] pays, quote, an ungodly amount of taxes
[173:40] to a government that now wants to spy on
[173:42] its own citizens and flatly stated that
[173:44] it will move its HQ out of the country
[173:46] if the bill passes.
[173:48] Even big tech.
[173:49] >> Apparently, this is directly supported
[173:52] by the current governing body. So, I got
[173:54] to take back what I said. My bad.
[173:56] >> Even big tech. Well, I I don't think you
[173:58] have to necessarily take back what you
[174:00] said. You did say with some exceptions.
[174:04] >> I guess this will be an extremely strong
[174:06] exception.
[174:07] >> Apple, Google, and Meta have all pushed
[174:10] back on the bill. And Apple has already
[174:12] shown its willingness to follow through
[174:14] on this kind of push back, having killed
[174:15] its encrypted cloud service in the UK
[174:17] over a similar sort of backdoor order.
[174:20] The public safety minister says
[174:22] amendments are coming to clarify that
[174:25] breaking encryption won't be allowed,
[174:28] but is holding firm on the one-year
[174:31] retention period. Our discussion
[174:33] question is, if privacy focused services
[174:36] actually exit Canada, does the bill end
[174:39] up making Canadians less safe by pushing
[174:41] them towards sketchier offshore
[174:43] alternatives, or is that an acceptable
[174:45] trade-off for giving police the tools
[174:47] they say they need?
[174:53] Uh, there's offshore alternatives that
[174:55] aren't that sketchy.
[174:57] Yeah,
[174:57] >> I think if people are using VPNs now in
[175:00] Canada, they will continue to be able to
[175:02] do so.
[175:03] >> If China with all of its technological
[175:06] wherewithal cannot figure out how to
[175:08] actually maintain the great firewall,
[175:12] don't get me wrong, Canada, love you.
[175:14] You're cool. Beavers, rocks and trees,
[175:17] modern stuff, very cool. I don't think
[175:20] you're gonna pull off um keeping
[175:23] offshore privacy tools out of Canada
[175:26] anyway. So, I'm I'm not convinced that
[175:30] this is going to make much of a
[175:32] difference. I could just continue to
[175:34] subscribe to Windscribe, who is now
[175:36] based in Europe somewhere.
[175:39] I don't know. I don't actually use
[175:41] Windscribe. I use PIA. But same same
[175:43] diff. I I'll just continue to subscribe
[175:45] to whatever service I already use in all
[175:48] likelihood.
[176:00] >> Norard says that this is deeply crappy,
[176:02] but would also depressingly only bring
[176:04] them up to parody with the Patriot Act.
[176:06] I actually don't know much about the
[176:08] Patriot Act. I haven't thought about it
[176:10] since like Snow actually. Has Snowden
[176:12] been in the news at all recently?
[176:15] >> I don't think so. I I looked him up
[176:17] recently actually. It seems like it
[176:26] >> Yeah, he's still living in Russia, it
[176:28] seems.
[176:29] >> Okay.
[176:30] Hasn't been uh drafted or anything.
[176:34] >> Yeah, I was going to say that's got to
[176:35] be an uncomfortable situation right now.
[176:40] Uh, what else we got here?
[176:43] AMD claims that its 256
[176:47] core Zen 6 Venice CPU beats Nvidia's
[176:51] Vera by 3.3x in rack level performance.
[176:55] They presented early projections showing
[176:58] that with both configured to the same
[177:01] 100,000W
[177:02] rack power envelope, they could
[177:04] dramatically outperform Nvidia's
[177:06] upcoming CPU. The comparisons based on
[177:08] estimated and modeled performance data,
[177:10] not completed side-by-side benchmarks of
[177:13] shipping products. And AMD describes the
[177:14] results as preliminary projections
[177:16] intended to indicate expected
[177:17] performance trends rather than final
[177:19] measured outcomes. Basically, this
[177:21] presentation was a response to Nvidia's
[177:24] own early Vera benchmarking campaign,
[177:26] which was based on Nvidia controlled
[177:28] testing. AMD's projections were built
[177:30] using methodology that was derived from
[177:33] those Vera results. And my notes go on,
[177:36] fight, fight, fight. Um, I mean, it's
[177:40] yeah, competition. It's cool. Always
[177:41] always fun to grab a grab a bucket of
[177:43] popcorn and watch. I do really feel that
[177:47] AMD might be maybe intentionally missing
[177:49] the point a little bit here. I don't
[177:51] think anybody is um I don't think
[177:54] anybody expects Vera to be like a top
[177:56] performing CPU. I think the whole point
[177:59] of it is just the the deep integration
[178:02] with uh the new Reuben GPUs. Um and I
[178:07] don't think Nvidia is trying to go for
[178:09] the general purpose CPU market, at least
[178:11] not at this time. But, uh, yeah. Hey,
[178:14] you you know what? Keep fighting AMD.
[178:17] >> Still kind of rooting for you even
[178:19] though you kind of really [ __ ] it up
[178:21] on the Thread Ripper side.
[178:25] >> Uh, me too, Dan.
[178:26] >> Speaking of
[178:28] >> Segways, uh, drone boat picked up downed
[178:31] US Army helicopter pilots, a first for
[178:34] sea rescues. After a US Army helicopter
[178:37] went down near the street of Hormuz
[178:38] while on patrol, its two pilots were
[178:40] rescued by a Corsair boat uh drone built
[178:45] by Seronic, a startup that makes
[178:47] autonomous surface vessels or ASVs.
[178:50] Using the vessel's 360° passive sensors,
[178:53] the Navy's Task Force 59 located the
[178:56] downed crew and picked them up within 2
[178:58] hours,
[179:00] bringing them to a safe spot on the
[179:02] water for a helicopter pickup. Important
[179:04] note, the Corsair's embedded AI software
[179:06] stack does the heavy lifting, handling
[179:08] immediate momentto- moment navigation.
[179:10] It dynamically plans paths around
[179:13] changing sea states and is supposedly
[179:15] able to continuously weight in position
[179:17] and regulate fuel and power consumption,
[179:20] only engaging the engine when needed.
[179:22] Peronic specs, the Corsair model is a 24
[179:26] ft ASV. We previously defined that, that
[179:29] can haul up to 1,000 lb across 1,000
[179:32] nautical miles. I really like that
[179:33] number. With a top speed of uh nope, not
[179:36] 1,000, 40 m per uh sorry,
[179:40] 40 mph. Uh Seronic has two other larger
[179:44] ASV models as well. And notably, this is
[179:46] believed to be the first unmanned sea
[179:49] rescue in US military history performed
[179:52] by boats that only entered into service
[179:54] in March. Imagine being
[179:57] the recipients of that. Um,
[180:01] yeah.
[180:03] >> I'm just looking for something. There's
[180:05] one last maybe topic here.
[180:11] Here we go.
[180:13] I'm going to forward this to Dan.
[180:19] Daniel Besser. There he is.
[180:23] All right. Before we get to After Dark,
[180:25] um Ashtech
[180:27] asked really nicely for me to do a bit
[180:30] of an update from them. So, if you're
[180:32] not familiar, Ashtech makes Hexos, which
[180:34] is the NAS software that I backed a
[180:37] little while ago.
[180:39] I did. It'll take a second.
[180:42] Um, so it's based on Trunaz, but it's
[180:45] like a a simplified front end slash also
[180:48] becoming a lot more than that, which is
[180:52] exactly what they uh wanted me to talk
[180:54] about. XOS lets you do so much more.
[180:58] Um,
[180:59] here are a few things. App preconfig.
[181:03] This feature is coming very soon in
[181:05] June, possibly next week. The first set
[181:07] of apps will be Plex, Jellyfin, and MBI.
[181:11] And what app preconfig is is basically
[181:14] like a one-click setup of apps like
[181:18] Plex. And they sent over a pretty cool
[181:20] little demo. Dan, do you have that
[181:22] hexosplex preconfig.mpp4?
[181:26] >> Did you get the email yet?
[181:29] >> Got him. Cool.
[181:32] >> Yes.
[181:33] >> Okay. Okay, so just to make it clear,
[181:35] this is never forced, but the goal is
[181:37] just to reduce the amount of steps
[181:39] required and cognitive load on the user
[181:40] to get a functional app experience,
[181:43] making apps far more applike and far
[181:46] less um
[181:49] figure out a whole bunch of crap like uh
[181:52] let me know when you're ready, Dan.
[181:57] >> Nope.
[181:59] >> Check this out, Luke.
[182:00] >> Ready?
[182:01] >> Yeah, go for it.
[182:03] There's a one minute time lapse, but
[182:05] basically this is the whole thing. So,
[182:07] see that preconfigure thing.
[182:12] >> Uh, so you sign in to your Plex because
[182:16] obviously like they can't sign in for
[182:18] you. You have to do that.
[182:20] >> Yeah, sure.
[182:21] >> Okay, there's your name and here we go.
[182:27] So, it obviously has to install the the
[182:29] Docker. And then on the left is the old
[182:32] way. This is actually a pretty cool
[182:34] little demo that they put together.
[182:38] Blah blah blah blah blah blah blah blah
[182:41] blah. Go find folders. Blah blah blah
[182:44] blah blah blah blah blah blah blah blah
[182:46] blah blah blah blah.
[182:48] Finish configuring all the folders.
[182:52] >> Are they going to do this for Jellyfin?
[182:53] Have they done this for Jellyfin
[182:54] already? Uh, Plex, Jellyfin,
[182:58] and MB will be the first set of apps
[183:00] with this preconfigure
[183:02] thing.
[183:03] >> Cool.
[183:06] >> Launch.
[183:10] So, if you're using their preconfigured
[183:12] folder structure, which the idea is that
[183:16] you will, there it is. Oh, I love that.
[183:20] That's a nice touch. That's a nice
[183:22] touch. We got the LT screwdrivers in the
[183:25] >> Yeah,
[183:26] >> coming soon.
[183:28] MB and Jellyfin out of the box linking
[183:30] between the R apps and Cubit Torrent and
[183:32] much more. Okay, so they're being pretty
[183:34] explicit about that. Okay, thank you for
[183:37] that. Um,
[183:41] John,
[183:41] >> hell yeah. Hell yeah.
[183:45] >> Okay, what else they got? Uh, additional
[183:47] app curations are also available now. So
[183:49] these are ones that are just like
[183:50] validated but don't have necessarily
[183:53] that oneclick setup. So flatter that's a
[183:55] jellyfin front end. Uh Jellystat
[183:57] statistics app for jellyfin and MB. Seir
[184:00] is a request management and media
[184:01] discovery tool that works for jellyfin
[184:03] plexer MB servers. That's a pretty funny
[184:07] uh thing to call that. Request
[184:09] management and media discovery. Okay.
[184:12] MKV tool mix is a set of tools to
[184:14] create, alter, and inspect metrosa
[184:16] files. Gee, I wonder why you would need
[184:18] that. And key is a fast and powerful
[184:21] bittor web UI. There's also more coming
[184:24] over the next several weeks. And they
[184:27] wanted me to talk about any raid. So
[184:29] there's a webinar on June 25th that if
[184:31] you're into like ZFS RAID, uh you might
[184:35] want to actually check out. So Alan Jude
[184:37] of Clara Inc. and uh John Penazo of
[184:40] Eztec, creator of Hexos, will be hosting
[184:41] a joint webinar to discuss ZFS any raid.
[184:44] So, how it works is it improves
[184:46] traditional it produ uh improves storage
[184:48] space efficiency. Um, and you can learn
[184:51] more on the Clara website about how it
[184:54] differs from traditional RAID Z. Also,
[184:57] buddy backups, you know how that's that
[184:59] was one of my key conditions for this
[185:01] thing being a thing that I will
[185:04] financially back. There's a timeline.
[185:07] The team has been focused on buddy
[185:09] backups since the 1.0 release. Um,
[185:12] things are going quicker than
[185:13] anticipated and they're hoping for Q3 of
[185:15] this year. So, sometime in the next
[185:17] three and a half months, buddy backups.
[185:20] So, I just want to explain what buddy
[185:21] backups are. Um, pretty much you have a
[185:24] buddy, you know, like this is my buddy.
[185:28] You both have a hexos NAS and you can
[185:31] both install some extra storage in each
[185:33] other's servers and without uploading to
[185:37] any cloud or going to anybody else's
[185:39] service, you can send an encrypted copy
[185:42] that is stored on your buddy's server.
[185:45] You can allocate that storage to each
[185:47] other so that you don't go over the
[185:48] quota that you set and you can have a
[185:51] secure safe off-site backup on hardware
[185:55] that you or your buddy controls and
[185:57] nobody else needs to look at or touch.
[185:59] And really the magic is not in doing
[186:01] that. That's something that you could do
[186:02] today uh with open- source free tools.
[186:06] You could totally set up something like
[186:07] that literally tonight in your spare
[186:09] time. Uh but the part that is a little
[186:12] bit harder is doing so without opening
[186:14] ports and without going through a bunch
[186:16] of networking tom foolery and command
[186:19] line configuration. So this is like
[186:21] click and go. Uh which is pretty pretty
[186:26] exciting. Uh Amaria says as long as my
[186:28] buddy can't see it. Yes, that is
[186:30] absolutely
[186:32] critical is that it would be encrypted
[186:35] such that only you know what you have
[186:37] backed up to your buddy's system. Uh,
[186:39] also they asked me to flag that a
[186:41] lifetime pricing increase is coming
[186:43] soon. The 199 price point is coming to
[186:45] an end. No exact date, but likely in Q3.
[186:49] Uh, they need to deploy a new e-commerce
[186:50] solution to replace the blah blah blah
[186:52] blah blah. Once that's done, lifetime
[186:53] will be going to 299. So, just watch out
[186:55] for that. All right, that's it. There
[186:57] you go, John. I did your I did your
[186:59] thing. It's legitimately really exciting
[187:01] stuff. I'm really stoked for Buddy
[187:02] backup and the app curations and the
[187:05] one-touch uh onetouch setup looks like
[187:09] honestly awesome. Like setting up Plex
[187:11] isn't even that hard for me now. I've
[187:12] done it enough times or excuse me, I've
[187:15] done it enough times that it's like
[187:17] fine. I know how to do it, but it's
[187:19] tedious.
[187:21] I'd rather not if I could just click one
[187:24] button and it's done. And all I have to
[187:26] do is just, you know, put my movies in
[187:28] my movies folder and put my pictures in
[187:30] my pictures folder and then it'll just
[187:32] do it for me. Yeah, that's that's pretty
[187:35] cool.
[187:36] Oh, shoot. Sammy demands float plane
[187:38] announcement. Oh, I'm so sorry, Sammy. I
[187:40] totally forgot about you.
[187:42] >> Before you do that,
[187:45] >> we're pointing out potential liability
[187:47] risks of the buddy backup thing. Um,
[187:52] >> we've talked about this in the past. If
[187:54] it's an encrypted copy, I just don't
[187:56] really see
[187:58] >> Yeah. If you if you genuinely truthfully
[188:01] have no way to access it and it's
[188:04] encrypted. I could see like if they end
[188:06] up getting investigated or something,
[188:07] they might like seize your drive.
[188:09] >> Yeah.
[188:10] >> Uh which would suck obviously.
[188:12] >> Yeah,
[188:13] >> they might seize your whole server.
[188:15] >> So you should probably only do this with
[188:18] buddies. But I mean by this logic um
[188:24] if you uploaded an encrypted file with
[188:27] incriminating
[188:29] media to Google Drive like
[188:34] is Google criminally liable for this? I
[188:36] don't think so.
[188:39] So am I missing something here?
[188:41] >> I I that's kind of where
[188:45] Yes, they are.
[188:47] They aren't. They would have no way of
[188:49] knowing if it's encrypted. Yeah. I I
[188:52] think I think people are not fully
[188:54] understanding what encrypted means a
[188:57] little bit.
[188:58] If it's an encrypted container file and
[189:02] you upload it to Google, they don't know
[189:04] what's in it.
[189:06] >> They're saying they they can be
[189:08] compelled to turn it over.
[189:10] >> Sure. Turn it over to what? Unless it
[189:12] can be decrypted,
[189:15] then it's then it's a it's a blob of
[189:18] ones and zeros.
[189:24] >> One second. Hello.
[189:27] Oh, thank you.
[189:29] >> So, he tried angrily complaining we
[189:31] haven't done float plane announcement.
[189:33] Now, he's trying buttering you up to do
[189:34] float plane announcement. Which one's
[189:36] going to work?
[189:36] >> He He is. Yeah, I think uh I think we
[189:39] should do the float plane announcement.
[189:40] >> I think we should. Thanks, Sammy.
[189:42] >> Have you ever wondered what our travel
[189:44] weeks look like? Sammy vlogged his
[189:46] entire week in Taiwan. From the
[189:49] immediate airport. I don't know what
[189:52] immediate Oh, from the immediate airport
[189:55] to ASUS shoot to I don't know what he's
[189:59] talking about. What a week LMG looks
[190:02] like. Wait, is this the one? No, this is
[190:06] an old one. Sammy, what the crap is this
[190:08] link? What are you talking about?
[190:11] You put the wrong link. I think he's
[190:13] editing it now. I see his cursor in
[190:14] here. Is that Sammy? No, that's Dan.
[190:16] Okay. I don't know. Whatever.
[190:19] So, he vlogged his entire week from
[190:22] airport to Isus shoot to going out for
[190:24] dinner with Taran and Luke who tried
[190:25] stinky tofu for the first time. So, if
[190:27] you're into how we get videos done, then
[190:29] this vlog is for you. Uh, but maybe
[190:32] you're more into Linux. Bam. We have
[190:35] some stuff that was taken out of the
[190:37] part 4 Linux challenge video. just ended
[190:39] up on the cutting room floor. So, we've
[190:41] got some extra from there. We will also
[190:43] have a video fully showcasing Lionus's
[190:45] motorcycle shooting soon, so let us know
[190:48] what you want to see from it. Lionus
[190:49] hasn't agreed to do this video, but I'm
[190:51] voluntelling him to do this video.
[190:53] LMG.ggan.
[190:55] Um, Ivonne did shoot me. Oh, I forgot to
[190:59] bring that. Oh, I forgot to bring that
[191:01] phone for Glenn today. Oh, crap. Uh,
[191:04] well, I will do it on Monday. Uh, Ivonne
[191:06] shot me going to pick it up, Sammy. So,
[191:08] you'll be able to use that for whatever
[191:09] it is that you're doing.
[191:12] Sammy says, "This is what happens when
[191:13] you write the doc at 7:00 a.m."
[191:16] >> I added in the link for you, Sam.
[191:18] >> Okay, here's the here's the proper link
[191:20] to the proper vlog.
[191:22] >> There you go. What a week at LMG looks
[191:25] like. Sammy POV Taiwan edition.
[191:31] What's going on in this?
[191:35] What are we looking at here? Oh, they're
[191:36] trying to show the fan working using
[191:38] like hair.
[191:40] I I eventually figured it out by just
[191:42] ripping a bigger clump of hair out of my
[191:44] head. So, if you if you watch that
[191:47] video, um they were like they were
[191:49] trying to shoot one hair and it was
[191:51] really hard. But, um I was like, "No,
[191:55] guys, you just need more hair." So, I
[191:56] just like like grabbed a whole thing of
[191:59] hair out of my head and then it worked
[192:00] pretty good
[192:03] here. Let me see if I can find it. It's
[192:05] a pretty cool video actually. If you
[192:07] guys haven't seen it, you definitely
[192:08] want to check out the ionic uh ionic
[192:11] cooling video. This perfectly silent fan
[192:13] took 300 years to make. Oh my god, ads.
[192:16] Well, we're I can't believe Roundup is
[192:19] legal again. That's crazy.
[192:22] >> Yeah,
[192:23] >> dude. I didn't really like understand
[192:25] what it was and I went a little ham with
[192:27] Roundup attacking weeds on our lawn when
[192:29] I was a teenager and like
[192:32] >> destroy emulated it. My mom was she came
[192:34] out then like a day or two days later
[192:36] like it h it acts really fast. She was
[192:38] like what the is going on every the lawn
[192:40] is dead.
[192:42] Yeah. Anyway,
[192:44] I was mostly a good kid.
[192:49] All right. Should we start a uh
[192:53] should we start after dark?
[192:54] >> After dark?
[192:55] >> Yeah.
[192:55] >> Yes.
[192:58] >> If I can find it. Maybe they didn't put
[193:01] it in the edit. Maybe it was too gross.
[193:03] If someone has a time stamp, let me
[193:05] know. Otherwise, I give up.
[193:11] Hit me, Dan. Sure thing. Are you guys
[193:13] excited for Legend of Zelda: Ocarina of
[193:17] Time remake officially being teased for
[193:19] later this year? And Luke, do you plan
[193:22] on finally having something to play on
[193:24] your Switch 2?
[193:28] The thing that is finally going to be
[193:30] playable on my Switch 2 is that game
[193:32] that is a remake from my childhood.
[193:34] >> Yeah, the game that already came out.
[193:36] >> Um, yeah, really. Uh, I had no idea that
[193:40] was coming. That actually
[193:43] does sound pretty cool. That was a sick
[193:45] game. I feel like it would still be
[193:47] really fun.
[193:50] >> Um,
[193:51] >> haven't played it in a billion years, so
[193:53] it might be fun to run through again.
[193:54] I'm sure I actually Sure, I remember a
[193:57] lot of it, but also don't remember a lot
[194:00] of it. So, yeah.
[194:03] >> I was not an Ocarina of Time kid. I
[194:05] didn't have an N64. My friend had an
[194:07] N64, but I um like I'm not like, but I'm
[194:11] going to go to my friend's house and I'm
[194:12] going to sit and play single player
[194:13] games.
[194:15] >> That's That's not That's not cool. So, I
[194:18] have tried to play Ocarina of Time
[194:22] on at least three occasions. Um, I think
[194:26] I actually own it for Virtual Console on
[194:28] my Wii and I tried it at least a couple
[194:31] times there and I tried it at least once
[194:33] emulated and I just
[194:36] um haven't managed to get into it. I
[194:40] have played through Link to the Past. I
[194:43] have played through most of The Legend
[194:44] of Zelda like the in the NES one. I've
[194:47] played through um Phantom Hourglass.
[194:50] I've played through uh Twilight
[194:53] Princess. I've played through Breath of
[194:54] the Wild. I'm like actually as far as
[194:57] game franchises go, I'm I'm as much into
[195:00] Zelda as I am into any franchise and I
[195:03] just
[195:05] couldn't Ocarina of Time. I'm so sorry
[195:08] to my fellow Zelda faithful.
[195:13] Question for Lionus. When Dbrand wants
[195:16] to do a sponsor with certain stunts,
[195:19] like uh the box of glass with a phone in
[195:21] it, do they tell the team or get any
[195:23] approval? Yeah, I genuinely typically
[195:27] don't know about it,
[195:31] which there's good things about that and
[195:34] there's also somewhat irritating things
[195:36] about that sometimes.
[195:40] Like there's been stuff that I've had to
[195:42] be like, "Yeah, no.
[195:46] Not very often, especially now that you
[195:49] know we understand each other pretty
[195:50] well. But um it has happened.
[195:54] So make of that what you will.
[195:57] >> Hello from Edmonton.
[195:59] >> Hi Edmonton.
[196:00] >> I drive a garbage truck. They have
[196:02] recently installed AI nanny cameras into
[196:05] our trucks. I find they encourage
[196:07] dangerous habits. I worry insurance
[196:10] could start demanding them. thoughts.
[196:14] >> Dangerous habits.
[196:17] I would love to hear more about what
[196:21] potential dangerous habits a nanny
[196:23] camera could encourage.
[196:27] >> Cuz I'm sure I believe you.
[196:29] >> That's gross. But yeah, I don't know.
[196:33] >> Like I genuinely believe you, but I
[196:35] don't I'm having a hard time imagining
[196:38] what that might be. I really wish I
[196:41] could have a little bit more context
[196:42] here.
[196:43] >> Maybe if you're operating too slowly, it
[196:46] gets mad at you.
[196:48] >> I could see that reaching production
[196:50] quotas.
[196:53] Um,
[196:55] also protos have increot increased this
[196:58] week. Get owned. Um,
[197:01] >> yeah, that makes that makes sense.
[197:04] >> I'm not sure though. I don't know.
[197:07] >> Oh, man. Okay, so
[197:10] this is Oh man. Okay, sorry. This is
[197:12] about to go a little bit off the rails
[197:14] here because my immediate knee-jerk
[197:16] reaction is like, "Yeah, nanny camera is
[197:17] terrible." Um, and then that made me
[197:22] think of like something that I was going
[197:24] to post as a tweet or something. I I was
[197:27] just going to do like a and obviously
[197:28] it's an observation that's been made
[197:30] many many times, but I wanted to post a
[197:32] picture of the like the pedal well of my
[197:35] Uber driver and there would be a brake
[197:37] pedal and then just like an onoff toggle
[197:40] switch on on the gas pedal. Um, and it
[197:45] it was I was thinking as I was sitting
[197:47] getting like violently nauseous in the
[197:50] back of a cabin or of an Uber in Taiwan,
[197:53] um, that ironically had one of those Did
[197:56] you ever see these those rule signs in
[197:57] the back of the Ubers there that will
[197:59] say like no smoking, no whatever, no
[198:01] throwing up. There's some that say no
[198:03] barfing. Um, it's like, bro, I wouldn't
[198:06] be choosing to vomit, but if you're
[198:08] going to drive like this, you're really
[198:10] putting me in a pretty awkward position.
[198:13] Um,
[198:14] anyway, it was it was occurring to me as
[198:16] I was sitting in the back of this Uber,
[198:18] not able to do anything other than stare
[198:20] out the front windcreen and hope that
[198:22] I'm not going to throw up that um it
[198:25] would be so easy for Uber to fix this
[198:28] problem by just using the [ __ ]
[198:31] accelerometer data that's already in the
[198:33] phone that the Uber driver absolutely
[198:35] has to have
[198:37] and just logging that and being like,
[198:39] "Hey, you need to chill the [ __ ] out on
[198:42] the brake pedal and the gas pedal so
[198:44] that your drive so that your passenger
[198:46] doesn't throw up in the back of your
[198:48] car. Like honestly, my driver to the
[198:51] airport was insane. And then I had at
[198:53] least two while I was in Taipei that
[198:56] just
[198:59] like crazy.
[199:01] >> Take a car to the airport.
[199:02] >> Sorry.
[199:03] >> Why' you take a car to the airport?
[199:05] >> Cuz it's cheaper.
[199:07] >> It's cheaper than the train.
[199:08] >> Yeah. If I'm gone for like 7 days. Wait,
[199:10] sorry. What are you asking? I'm supposed
[199:12] to take the train to the airport
[199:15] >> from my house.
[199:15] >> Oh, I thought you meant Sorry. Sorry. I
[199:17] thought you meant from the hotel in
[199:18] Taipei to the airport.
[199:19] >> No, no, no, no, no. From No, from V in
[199:21] Vancouver.
[199:22] >> Okay. Yeah. No, I understand. Okay.
[199:24] >> Yeah. So, once I'm once I'm gone for
[199:26] like I think I think the threshold is 5
[199:27] days or something like that. If I'm gone
[199:29] for more than 5 days, then it's cheaper
[199:31] to
[199:32] >> being so expensive. No, I got you. I
[199:33] thought I thought you Yeah, my um
[199:36] anyway. So, yeah. So, you've had you
[199:37] have the GeForce data from the phone.
[199:39] And then also another thing that just
[199:41] triggered the crap out of me is I went
[199:43] to leave like a three-st star review
[199:45] because realistically, we didn't crash,
[199:47] you know, they didn't like kill me, but
[199:49] also I was feeling really sick. And you
[199:52] know how it prompts you? It's like,
[199:53] okay, well, what was the issue?
[199:55] Nausea is not in there.
[199:59] How is it that with such a like
[200:01] universal meme, Uber drivers and their
[200:05] ridiculous acceleration and braking, how
[200:08] is it that that is not in there as an
[200:11] issue with your ride? That's crazy. It
[200:14] has to be intentional. There is no way
[200:18] that that's not intentional.
[200:20] It's It's weird that that's a known
[200:22] thing because you'd think that would be
[200:25] worse for gas and worse for car
[200:28] maintenance.
[200:32] >> I don't get it, dude.
[200:33] >> I don't know. Yeah, I don't know.
[200:34] >> It doesn't seem that complicated to make
[200:37] small adjustments to the gas pedal. I
[200:39] mean, and maybe part of it is just that
[200:41] I've been married to someone with like
[200:43] serious motion sickness issues for a
[200:46] thousand years. Um, and so I just I just
[200:50] automatically adjust my driving style
[200:53] when I have a when I have a passenger.
[200:55] Like when I'm in my car on my own,
[200:56] whatever. I don't care. I'll just open
[200:57] it up, right? Like it who cares? It's
[200:59] me. I'm not going to get sick driving my
[201:01] own car. But when I have a passenger,
[201:03] I'm like like trying to be smooth about
[201:06] it. And I guess maybe they just like
[201:08] never had anybody tell them. I don't
[201:10] know. I don't know.
[201:14] It just seems like they they do care
[201:17] about a lot of passenger issues. Why is
[201:19] nausea not one of them when it's just
[201:21] such a meme?
[201:27] My son listens to the show with me and
[201:30] your videos have him interested in tech
[201:32] and programming, but I worry about him
[201:34] watching other online content from tech
[201:37] bros. Any suggestions for young
[201:39] enthusiasts? Man, this is one of those
[201:42] classic questions people ask me that I
[201:44] just really don't know the answer to
[201:45] because the truth is I just don't
[201:47] consume a lot of YouTube.
[201:51] Sorry, Wendell. Wendell's like the most
[201:54] wholesome person that I know. Level one
[201:56] text is a little technical, but like
[202:00] >> he also explains things really well.
[202:03] >> Yes.
[202:04] >> I figure if you watch enough level one
[202:06] text, eventually you'll have some idea
[202:07] what the heck he's talking about. This
[202:09] was a young enthusiast of of what? Linux
[202:11] was it?
[202:13] >> No, just tech and programming.
[202:16] >> Tech and programming.
[202:23] I don't think Wendle goes over
[202:26] programming much,
[202:28] >> but tech.
[202:29] >> Tech definitely. Yeah.
[202:32] Yeah.
[202:34] I think a lot of the
[202:37] YouTubers I watched when I was trying to
[202:39] learn programming, I don't know if
[202:41] they're even still around. Um,
[202:44] let me see if I can
[202:48] if they are still.
[202:52] Maxwell in float plane chat says
[202:54] Hacksmith, Mark Robber, Michael Reeves,
[202:58] the one YouTube shorts guy. Is that what
[203:00] they're actually called or are you just
[203:03] memeing?
[203:05] Oh,
[203:06] >> the stuff that he just mentioned is is
[203:08] very fun on the entertainment side of
[203:09] things, but not necessarily the
[203:10] education side of things.
[203:11] >> Technology connections. Sure. Oh, yeah.
[203:14] Bent Bob says Lori Wired. That's that's
[203:16] that's a cool place to hang out.
[203:18] >> Definitely. That one would be really
[203:20] good, especially for inspiration.
[203:21] >> Yeah. Okay. There's a few there's a few
[203:23] good ones.
[203:25] >> I haven't watched this channel since I
[203:27] went to BCIT.
[203:30] So, it's been a hot minute, but I used
[203:33] to watch uh the New Boston for Java
[203:37] tutorials
[203:39] a million years ago.
[203:40] >> Oh my god, that's a name I haven't heard
[203:42] in a thousand years.
[203:44] >> Do you recognize that name?
[203:45] >> Yes.
[203:46] >> Yeah.
[203:47] >> Oh, that hurts.
[203:48] >> Um,
[203:50] yeah. Java programming tutorial one,
[203:52] installing the SDK from 17 years ago has
[203:54] 6.8 million views.
[203:57] >> Oh, Anie. Hey, Cash has another really
[203:59] good one. Branch education is amazing.
[204:02] >> Oh, branch education is sick.
[204:04] >> Yeah,
[204:05] >> that's a that's a great one.
[204:06] >> That should be enough to keep them busy
[204:07] until it's time for you to order
[204:08] something else from the store.
[204:09] >> Yeah, I like it.
[204:12] >> All right, hit me down.
[204:13] >> Hey, Luke and Linus. My boss is the
[204:15] executive at my work and I spend a lot
[204:18] of time ensuring he gets all his tasks
[204:20] done. As executives yourselves, do you
[204:23] have any insight how I can manage this
[204:25] better?
[204:27] Vance, Vance, are you still here?
[204:29] >> Yeah, you're asking the wrong person.
[204:31] Um,
[204:34] >> uh,
[204:36] I mean, making sure that the calendar is
[204:39] like set up with realistic amounts of
[204:43] time for things, I think, is really
[204:44] helpful. Um,
[204:45] >> yeah, consider travel time when booking
[204:47] things.
[204:48] >> Yeah, that's really helpful. Um,
[204:50] >> try not to actually book things exactly
[204:52] back to back all the time because
[204:54] sometimes the brain needs a little bit
[204:56] of a reset or you need to grab a water
[204:58] or go to the washroom or there's travel
[205:00] time involved or something. Um, I find a
[205:03] lot of my days I end up like pushing
[205:05] meetings back just constantly to the
[205:07] point where it's like almost a meme. And
[205:09] one of the reasons why is because I'll
[205:11] have like six hours straight
[205:15] of meetings with zero breaks between
[205:17] them that often do involve some form of
[205:20] like travel time or something else.
[205:21] >> And all of this is quite literally
[205:24] impossible. Like I Yeah.
[205:26] >> Um be be cognizant of decision fatigue.
[205:29] I'm going to use the horrible AI
[205:31] overview. Here we go. A state of mental
[205:33] exhaustion caused by making repeated
[205:35] choices throughout the day. Um, and
[205:39] every decision contributes to it. It's
[205:42] It's actually really interesting. It was
[205:43] something that I experienced for a lot
[205:45] longer than I knew what it was. But
[205:48] like, it was always very obvious to me
[205:51] that I should not think about my
[205:53] wardrobe at all in the morning. I should
[205:54] simply grab whatever t-shirt is at the
[205:56] top of my pile and whatever pants are at
[205:58] the top of my pile and just put them on
[206:00] because I don't have room in my brain to
[206:03] to think about that. and then also all
[206:05] the other things that I need to think
[206:07] about in a given day. Um, and so it it
[206:10] manifests in like weird ways like when
[206:11] I'm having a really intense morning
[206:14] already
[206:15] like Vance will ask me like hey do you
[206:17] want me to get something for you and for
[206:19] lunch and I'll be like uh and literally
[206:23] I can't think of what to eat. Um, and so
[206:25] that's just it's like it's a obviously a
[206:27] super first world problem, but finding
[206:31] ways to manage out
[206:34] pointless decisions that don't matter um
[206:38] so that they have mental capacity left
[206:42] for things that actually do have an
[206:43] impact on the business is something that
[206:45] I would say would be helpful.
[206:47] >> Yeah.
[206:50] >> Hi, Lionus. As a fellow lover of soup
[206:53] dumplings, I saw you went to Dintai at
[206:56] Computex. I've only been to one in the
[206:59] US, and I'm wondering if it's the same
[207:01] in North America. Have you ever tried it
[207:03] in Canada?
[207:04] >> I haven't. I think they mean California.
[207:06] Um I
[207:08] >> We actually have one.
[207:09] >> We have one.
[207:09] >> One.
[207:10] >> Yeah.
[207:11] >> Where?
[207:12] >> Uh I I looked earlier. Alurnie.
[207:16] >> Port Alurnie has a dent. No, the street.
[207:19] No, Alurnie Street. Oh, it's in
[207:20] Vancouver.
[207:22] >> Oh, that's hilarious. Vancouver.
[207:24] >> It is on Old Bernie Street. Yeah. Is
[207:26] that the same one?
[207:29] >> Is it like a change?
[207:30] >> It's the same company.
[207:32] >> It's a Yeah.
[207:33] >> locations.
[207:34] >> As far as my understanding goes, it is.
[207:35] Yeah.
[207:36] >> It's the same as all the other North
[207:39] America ones,
[207:41] but I don't see any actual
[207:45] Taiwan locations
[207:48] here.
[207:50] Dintyong North America. Okay, I don't
[207:53] actually know if this is the same
[207:54] company. Um, my my ah man, this is Oh,
[207:58] dude. I'm going to get cancelled for
[208:00] this for sure. Dentong is overrated.
[208:03] >> There, I said it. Oh, I was excited to
[208:05] try it. I love dumplings.
[208:06] >> It's No, it's I didn't say it's bad.
[208:09] >> By all means, try it. But there's
[208:10] >> It's not the best thing ever. It's the
[208:11] worst. Obviously.
[208:13] >> No. Is it overrated? Is it overrated?
[208:16] Because you've probably, I'm not
[208:18] exaggerating at all, had it 50 plus
[208:21] times.
[208:22] >> No, I've only been to Dintyang, I think,
[208:24] twice.
[208:26] >> Ever?
[208:27] >> Yeah.
[208:28] >> There's no way. I've been with you to
[208:29] Daang more than two times.
[208:31] >> Okay. Every time I've gone, I was with
[208:32] you then. Like, I've never gone there of
[208:34] like my own accord.
[208:36] >> Okay.
[208:37] >> So, like three times then.
[208:39] >> Um, almost like the soup dumplings at
[208:42] the hole in the wall down the road from
[208:44] our hotel are better.
[208:46] It's just straight up better.
[208:48] >> No, that's true. That is a really good
[208:49] spot.
[208:49] >> Like I just I've
[208:51] >> I have had
[208:53] Xiaoong Bao and other soup dumplings
[208:56] >> at a variety of other places in Taiwan
[208:58] that are worse.
[208:59] >> And Taiong,
[209:01] >> one of the problems with Dintyong is
[209:02] that it it is not the best. It's really
[209:06] solid and it's very consistent, but it's
[209:08] not the best.
[209:10] >> Very expensive compared to pretty much
[209:12] everywhere else.
[209:13] >> So expensive. Like I pay practically
[209:15] nothing for those ones down the street.
[209:17] I watch them make them. They taste
[209:20] freaking incredible. And then Dintyong
[209:23] is like it's fine.
[209:25] >> And to be clear, I didn't say bad. I
[209:28] just said overrated
[209:31] >> because people just like cream their
[209:33] freaking pants talking about how amazing
[209:35] Dint Taiong is. And I'm like, "Yeah,
[209:38] it's good, but like
[209:41] it better be the best thing ever at that
[209:43] price and with that reputation and with
[209:46] that long of a wait to get into it."
[209:48] >> Yeah. Like I I know people say like,
[209:50] "Oh, if you're coming to Taipei, you
[209:51] have to get Dint Taiong."
[209:54] >> Say, "No, if you're going to Taipei, you
[209:56] should
[209:57] >> get really good soup dumplings."
[209:59] >> Yes.
[210:00] >> Um some really good xiaolong bao.
[210:03] >> Yeah. And like I've had better beef
[210:05] noodles. places you can get that.
[210:06] >> I've had better beef noodles like many
[210:08] many times. I would say Dint Taiung is
[210:10] like a solid B minus for beef noodles.
[210:14] >> Yeah, that sounds about right.
[210:15] >> Like it's
[210:16] >> everything there other than their their
[210:18] dumplings are kind of whatever.
[210:19] >> It's fine.
[210:21] >> Yeah. It's also incredibly consistent. I
[210:25] I don't know if I've been basically
[210:27] anywhere that's as consistent as Dinton
[210:29] is, which makes it pretty good. It's
[210:34] also it's highly unoffensive and it's
[210:36] extremely consistent. So, it makes it
[210:38] pretty good for like meetings.
[210:39] >> Yeah, it's kind of westerner friendly.
[210:41] Like they have something to eat for the
[210:43] people in your party.
[210:44] >> Basically, everyone
[210:45] >> don't eat seafood or you know, whatever.
[210:47] >> It's one of the easiest things to be
[210:49] like, "Oh, we have like 10 people and we
[210:51] all need to get dinner. What is a place
[210:52] we can all agree on?" And usually that
[210:55] place does work.
[210:56] >> Yeah.
[210:57] >> Um
[210:59] >> Yeah.
[211:02] Question for whoever feels like
[211:03] answering. Do you guys ever think we
[211:06] will have a solution for kernel level
[211:07] antiche on Linux? If yes, what can it
[211:10] be? Will I ever see Lol or BF running on
[211:13] Arch? By the way, okay, I got an
[211:17] absolute wall of text about our
[211:20] discussion
[211:22] regarding anti-che on Linux a couple of
[211:26] weeks ago.
[211:28] >> Whoa. from someone who was very upset
[211:30] about like seriously I'm still going.
[211:34] Um,
[211:36] and
[211:37] >> I got I got some of my stuff taken
[211:38] pretty dramatically out of context on
[211:40] that topic. Um basically the take was
[211:45] there's really good tools in Linux for
[211:48] managing all of this and it wouldn't be
[211:50] necessary to have kernel level antiche
[211:54] on Linux because of like how how well
[211:58] you know managed it is
[212:02] but I acknowledge that no current DRO
[212:06] has actually like you know implemented
[212:09] this level of of of lockdown that would
[212:12] give developers the confidence to to do
[212:16] it. So it's maybe something for it to
[212:20] figure out later or something was I I
[212:22] read the whole thing but it was like a
[212:25] week ago or something like that. When
[212:26] did I read this? Yeah, I read this about
[212:27] a week ago and I was just like okay. So
[212:30] basically nothing that Luke or I said
[212:32] about the unlikelihood of this happening
[212:35] due to the very open nature of Linux is
[212:38] actually untrue. Um because
[212:42] the Linux crowd is very like, you know,
[212:45] own your hardware, do whatever you want
[212:47] with your hardware. Oh, by the way, also
[212:49] your software, too. The likelihood of
[212:52] this kind of a dro existing
[212:55] seems very
[212:57] low.
[212:59] Um
[213:02] yeah, it's I don't know. the the
[213:04] cheating situation is
[213:08] really really bad right now. Um,
[213:13] all shooter games are rampant with it.
[213:17] Usually when shooter games have an open
[213:19] beta for a weekend, people are found
[213:22] rage cheating
[213:25] on that one weekend. Like, it's it's
[213:27] it's wild. Shooter games are basically
[213:29] cooked. Um,
[213:32] Counter-Strike lobbies are just riddled
[213:34] with cheaters all the time. There was a
[213:36] massive bandwave recently. Um, I was all
[213:40] pissed over Tarov years ago. Um, and
[213:45] then lo and behold, it just turned out
[213:47] it's every shooter game just has
[213:49] cheaters everywhere. Um, they're also
[213:52] highly prevalent in like basically every
[213:54] other game. Um, cheating is really
[213:57] brutal and really rampant right now. And
[213:58] it's honestly like ruining competitive
[214:01] multiplayer gaming. Um,
[214:04] >> are you sure you're not just unk?
[214:07] >> Actually, yes. Um, the cheating
[214:10] situation is way worse than ever. The
[214:12] reason why the cheating situation is way
[214:13] worse than ever is because money is
[214:15] highly involved now. Um,
[214:17] >> there's so much money which is crazy.
[214:19] The the reason the the only reason why
[214:21] I'm sure not just UNC to be clear cuz I
[214:23] could be um is because I watch people
[214:26] like Shroud I think I talked about this
[214:28] on W show before Shroud went back and
[214:30] played Counterstrike on Shroud's not
[214:32] servers.
[214:33] >> Shroud is not Shroud is still a god.
[214:36] >> I'm just kidding. I'm just kidding.
[214:38] >> When when Shroud stops being a god, I'll
[214:40] I'll probably be dead. Um but
[214:44] Shroud went back and played
[214:45] Counter-Strike on normal servers and
[214:46] there was a cheater in like literally
[214:47] every lobby he was in. like it's
[214:49] actually just it's insane. So, it's a
[214:53] touchy subject and having
[214:58] worse
[215:02] anti-che on Linux when the situation is
[215:04] as bad as it is um
[215:08] can can suck for some people, especially
[215:11] when it's an like an uneven playing
[215:13] field of anti-che, if that makes sense.
[215:16] where you can play the game on Windows
[215:19] or Linux, but the Linux one has like
[215:22] less anti-che than the Windows one does.
[215:26] Um,
[215:28] which could bring a bad reputation to
[215:30] Linux because then there's a higher
[215:31] chance uh a higher perceived chance that
[215:36] the cheaters might be running Linux. Um,
[215:40] I don't know. It's a tough situation. I
[215:42] also am highly acutely aware that these
[215:47] kernel level anti-cheats exist in the
[215:49] environment of oh my god we've never
[215:51] seen this many cheaters before. So
[215:53] clearly they're not working anyways. Uh
[215:56] it's it's a tough spot. Um I think
[215:59] there's a lot of people that don't want
[216:01] kernel level anti-che coming to Linux
[216:03] and I can sure understand that. That
[216:05] would blow. I don't know how we solved
[216:08] the cheating problem.
[216:11] Um,
[216:13] I know that right now the main way that
[216:16] anti-che systems are getting on Linux is
[216:19] they're just more lax than they are on
[216:21] Windows. At least I believe that's
[216:23] what's happening. Um,
[216:28] again, I don't really know if it matters
[216:30] that much if
[216:32] most of these cheats are getting around
[216:34] these systems anyways.
[216:36] So,
[216:38] >> just to clarify my from my wall of text,
[216:41] uh, nothing would prevent a regular DRO
[216:43] and game developers from implementing
[216:45] it. But with all that context, and there
[216:48] was a lot, I'd say you might even be
[216:50] right in that who knows if companies
[216:52] will want to put in the time to make
[216:53] these things work. Um,
[216:59] yeah.
[217:01] So,
[217:03] I
[217:05] Yeah I
[217:06] Captain Immortal is saying I think
[217:08] cheaters al sorry I think we've got some
[217:09] lag issues. I think cheaters are also
[217:11] spoofing being on Linux to get a lower
[217:13] anti-che even when they're on Windows.
[217:15] Um I don't Yeah, I don't know. Um it's
[217:19] just again I don't even know if it
[217:20] matters. Like we have we have this
[217:22] insane cheating problem
[217:25] and we have kernel level antiche. So
[217:27] it's like
[217:29] I don't know if it helps. I was reading
[217:31] an article recently saying it's going to
[217:33] move like even higher up the
[217:37] I I forget what even space we're in
[217:39] above kernel
[217:42] um lower.
[217:45] >> Hold on. I there's no way I'm going to
[217:47] be able to find this article right now.
[217:50] Forget it.
[217:51] >> Yeah, Denrix is saying BF6 had like a
[217:53] day one of beta cheat. Um there's
[217:56] someone else saying that Arc Raiders has
[217:58] a rampant cheating problem right now.
[217:59] It's what I'm finding is with basically
[218:02] every shooter game, you have to wait for
[218:04] like with Tarov, you have to basically
[218:06] wait for a wipe and then you can play
[218:07] for a couple months and then it gets too
[218:09] bad. Um, with other games, you have to
[218:11] wait for like a ban wave or some major
[218:14] release and then you can play for a
[218:15] little bit and then it will get too bad.
[218:17] >> Um,
[218:19] it's like the the the gaps between banan
[218:22] waves
[218:24] are basically just unplayable. Um, and
[218:28] and it's kind of interesting cuz it's
[218:30] making this like cycle of firstp person
[218:33] shooter games where people will jump
[218:34] from game to game in between these banan
[218:37] waves and like what an insane situation.
[218:40] Um, Griffon says firmware anti-che.
[218:43] Yeah, I think it was like at the
[218:44] firmware level, but I forget how they
[218:46] were trying to implement that. I forget
[218:48] what that was supposed to look like. Um,
[218:51] and I don't want to get anything wrong,
[218:52] but it was like it was like even deeper
[218:55] embedded into the system. And I just I
[218:57] forget what um what the kind of plan
[219:00] was, what the direction we were going
[219:02] was. And I mean, again, all of this
[219:06] is sort of a it's an it's an it's a it's
[219:08] a it's an arms race that can never be
[219:10] won because in the age of machine vision
[219:14] and robotic arms, which is coming, you
[219:17] will be able to cheat without any
[219:21] software installed on the system at all.
[219:26] At some point, you know, if you build a
[219:28] perfect antiche,
[219:30] someone will build a robot arm that
[219:33] emulates the slight, you know, movement
[219:36] of a human hand, but then just moves as
[219:40] quickly as possible to I don't
[219:42] understand why people even want this. As
[219:44] someone who just plays games for the fun
[219:47] of actually competing against people, I
[219:49] don't understand why people are paying
[219:51] for this. But if someone will pay for a
[219:53] software that moves the crosshair to
[219:56] their opponent, then I see no reason why
[219:58] they wouldn't also pay for a hardware
[220:00] that moves their mouse across their
[220:01] mouse pad that moves the crosshair to
[220:04] the opponent. It seems
[220:05] >> and they will play they will pay a ton
[220:08] of money for that thing that you don't
[220:11] understand why they would pay for it.
[220:13] Some of these cheats are incredibly
[220:14] expensive and subscription basis as
[220:17] well.
[220:18] Um, that was uncovered when people were
[220:21] diving into the Tarov stuff. People are
[220:23] paying hundreds of dollars a month.
[220:26] >> Charge Nuclei says, "Pay a better player
[220:30] to play the game for you." I mean,
[220:32] >> only trillionaires can afford that sort
[220:34] of thing.
[220:35] >> Yeah. Yeah. Yeah. Yeah.
[220:36] >> Yeah.
[220:39] >> Yeah. That also happens.
[220:42] >> Uh, some more for you.
[220:44] >> Yeah. Hit me.
[220:46] >> Hi. Bad cholesterol. Luke, any cool
[220:49] techy tours you've been on that you can
[220:51] talk about
[220:54] techy tours?
[220:57] H.
[221:01] Yes, absolutely. I don't know how much I
[221:03] can say about them. We got to see um
[221:07] Intel's performance testing lab.
[221:09] >> Mhm.
[221:10] >> That was really cool. We got to see some
[221:13] stuff at Qualcomm. That was really cool.
[221:16] We got to see
[221:19] more recently
[221:21] uh some of the people from the lab went
[221:23] down to Puget Systems to see some of
[221:26] their like lab stuff similar to ours and
[221:30] talk shop for a while. That was really
[221:31] awesome. It was interesting to see
[221:35] kind of the differences
[221:37] uh because they have they have a
[221:39] different use case for their lab than
[221:40] than ours. um they're mostly interested
[221:44] in in testing hardware that they will be
[221:48] selling if that makes sense. So there's
[221:50] kind of a narrower scope of the things
[221:52] that they need to test. They're also
[221:55] testing like patches of things u to see
[221:57] if anything major changes or if things
[221:59] break or whatever. It's very
[222:01] interesting. Uh different goals but very
[222:03] interesting. Um
[222:08] what can I really say about them? It's
[222:10] been very surprising to me.
[222:17] Got to frame this in the right way.
[222:24] Yeah, it's been very surprising to me
[222:26] how close tech YouTubers are to some of
[222:31] these really big companies in regards to
[222:34] their testing capabilities.
[222:36] >> Oh, yeah.
[222:37] It's I I didn't realize how advanced the
[222:41] tech YouTube space was. To be completely
[222:44] honest, I was expecting walking into
[222:46] some of these big companies, seeing
[222:47] their performance testing labs and
[222:49] evaluation labs and like having my brain
[222:51] explode. And I walked in there and was
[222:53] like, "Yeah, okay. I mean, I recognize
[222:56] most of this." There was some weird
[222:58] stuff. I'm I'm going to start getting
[222:59] more vague about where we were cuz I'm
[223:01] talking about areas that I wasn't able
[223:02] to bring a camera. Um, but there was
[223:05] some stuff that got into like the radio
[223:07] side of things, um, like wireless
[223:10] technologies that got a little like, oh,
[223:13] okay, that's interesting. Um,
[223:17] >> and testing of like cameras and camera
[223:20] software and stuff that that that also
[223:23] got pretty outside of the the wheelhouse
[223:26] of what I've seen from from obviously us
[223:29] included, but tech YouTubers in general.
[223:31] Um, but in in regards to like standard
[223:34] performance.
[223:36] >> Yep.
[223:37] >> Yeah.
[223:37] >> As soon as you walk into like a failure
[223:39] analysis lab, you're going to see single
[223:41] pieces of equipment that are worth more
[223:43] than our entire lab, including the
[223:45] building.
[223:46] >> But
[223:47] >> yeah, like we we ain't doing that.
[223:48] There's there's no stuff. But when we're
[223:50] talking about performance testing,
[223:51] >> measuring frame rates,
[223:53] >> yeah,
[223:54] >> they're at the mercy of
[223:55] >> they're at the mercy of software
[223:57] providers like um uh Maxon the same way
[224:01] that we are to provide a usable
[224:03] benchmark. Um yeah, it's an industrywide
[224:07] challenge.
[224:08] >> Yeah, it's interesting.
[224:12] >> Message to Lionus. How happy were you
[224:15] when Apple admitted that iOS 26 was slow
[224:18] and buggy? Uh, proving that it wasn't a
[224:21] skill issue.
[224:21] >> Listen, I knew it wasn't a skill issue
[224:24] already because like I didn't do
[224:26] anything and it would just have all
[224:27] these problems. Um, but I I I don't
[224:32] know. The Apple faithful who just like
[224:35] glaze them endlessly are never going to
[224:37] even when Apple themselves says it,
[224:39] they'll find a way to excuse it. Um, I
[224:43] wasn't even giving Apple a particularly
[224:45] hard time about it. I basically was just
[224:47] like, I expect more from Apple. This is
[224:49] as bad as like a buggy Android release
[224:51] and I got like crucified for it. Uh, so
[224:54] yeah, I know I I knew I was right
[224:56] because I had so many bad experiences
[224:59] with it and Ivonne is like fed up at
[225:01] this point. She asked me to get her an
[225:04] Android. We almost went shopping for it
[225:05] in Taiwan and then the tech mall that we
[225:07] were at didn't have any phone stores.
[225:09] So, we we were going to just shoot an
[225:10] impromptu video of me buying a phone for
[225:12] her, but um there there were no stores.
[225:18] >> Man, she might like the Xperia.
[225:20] >> Dude, I hate buying no Experia for Do
[225:23] you see how much those cost? I could buy
[225:25] her jewelry for that kind of money.
[225:27] >> Oh, yeah. They are crazy expensive,
[225:28] aren't they? What's up with that, man?
[225:30] Damn,
[225:31] >> they're good. Shut. Uh
[225:34] >> so expensive, though. That's why you buy
[225:36] last year's model or get one from a
[225:38] coworker.
[225:40] I still haven't I still haven't wiped
[225:42] it. I got it caught up cuz my Steam
[225:45] guard was still on the old one. And if I
[225:47] hadn't have done that, it would have
[225:48] been absolutely hooked.
[225:49] >> Oh jeez.
[225:50] >> EDM. No, there is no difference between
[225:52] high-end and low-end Apple devices. And
[225:54] I'm not definitely right about the lower
[225:55] end of things. I had a freaking the
[225:57] latest iPhone. What are you talking
[225:58] about?
[225:59] >> This is what I'm talking about.
[226:01] >> Stop. Just stop engaging.
[226:03] >> Yeah. Move on.
[226:04] >> Move on. Next one.
[226:05] >> Hi, Ll and D. The IT department of the
[226:09] school I work at is restricting the use
[226:11] of Muse Score because it's open source.
[226:13] How would you explain or convince
[226:15] someone the safety of a longestablished
[226:18] project?
[226:18] >> Point at their phone. If they run an
[226:20] Android, you say, "Hey, that's based on
[226:22] the AOSP, an open- source project. You
[226:24] need to relax."
[226:26] >> Like, what are they talking about?
[226:29] >> They are in the pocket of Selius.
[226:34] What's Musecore is? I'm assuming this is
[226:36] some music thing.
[226:36] >> Yeah, it's a music notation software.
[226:39] >> Okay,
[226:41] they banded sheet music.
[226:42] >> That is the worst, dude.
[226:45] >> Scores.
[226:45] >> What are we talking about?
[226:46] >> Pretty dang.
[226:47] >> Tell them to ask an AI.
[226:49] >> Wait, this is the IT department. They
[226:51] should know better.
[226:52] >> That's what makes no no bothering me.
[226:55] >> No, no, no, no, no, no, no. there's some
[226:57] other reason that is not the real reason
[227:00] and they're just not telling you the
[227:02] real reason. So you may never be able to
[227:05] find out the real reason, but until you
[227:07] know the real reason, you won't be able
[227:09] to make any argument or explain
[227:11] anything. So that's that's that's the
[227:14] situation you're in right now. There's
[227:15] no way that an IT department at a school
[227:19] who knows enough about it to be running
[227:21] the IT department at a school
[227:24] >> does not know about open source. Like
[227:27] there's just no way.
[227:28] >> I was
[227:30] I I not knowing what MuseCore was. I I
[227:33] thought I misinterpreted that for a
[227:34] second and they they banned the use of
[227:36] Musecore because there was some other
[227:37] opensource one that they wanted you to
[227:40] use and that would have made like more
[227:42] sense to me as a educational
[227:45] institution. But
[227:47] my school IT department gave us all
[227:49] local admin.
[227:52] That's fun. Maybe they just wanted you
[227:54] to learn things rapidly.
[227:59] Uh yeah. Okay. Next one. Hello LDL
[228:03] watching today while painting my nursery
[228:05] for my first kid do this month. Any DIY
[228:09] painter tips slash first time dad
[228:11] advice.
[228:12] >> Well, don't mix up the paint with the
[228:14] babies.
[228:16] Babies and paint are not a good
[228:17] combination. So, I think it's great that
[228:20] you're, you know, Oh, due this month.
[228:22] Oh, dude. Uh,
[228:26] you needed to paint this like a month
[228:27] ago. The fumes are not going to be good.
[228:30] You need to get on it, finish it this
[228:32] weekend, and then run a fan in there
[228:35] like continuously. Don't put the baby in
[228:38] there for a little bit. Um, so there
[228:40] there's one. I got one for you. There's
[228:43] a tip. Um,
[228:46] what else we got here?
[228:49] >> Yeah, good enough.
[228:51] >> Yeah. Hello, L D and lowercase L. with
[228:55] how easy AI generation of creators has
[228:58] become, what advice do you have for fem
[229:00] creators to protect themselves?
[229:03] >> Uh,
[229:04] I mean, I wish I could offer some advice
[229:06] on that. I don't know how to protect
[229:08] ourselves. I I think I I think I
[229:11] mentioned this last WAN show, but I was
[229:12] chatting with an automotive creator and
[229:15] uh they were saying like, oh yeah, no,
[229:17] like AI generated slop is like a real
[229:19] challenge for our space. And I was like,
[229:21] sorry, why? Um, and they were basically
[229:24] just like because you can just generate
[229:26] cars doing like crazy stunts and people
[229:30] the same at the end of the day it's zero
[229:32] sum, right? Like people have a finite
[229:34] amount of eyeball time per day and every
[229:39] minute they spend looking at AI slop of,
[229:42] you know,
[229:45] a a goat car, you know, drifting around
[229:48] a supermodel or whatever is time that
[229:50] they're not spending looking at the
[229:52] content that was made by real people
[229:54] with, you know, real cars or whatever.
[229:56] And uh it was something that I hadn't
[229:58] really
[230:01] thought about enough like I yeah like
[230:03] I'm aware of the concept of that like
[230:06] when the Super Bowl is on actually
[230:08] YouTube viewership goes down you know
[230:09] like I know these things but the scale
[230:12] at which this just sort of generated
[230:14] garbage content is being produced now is
[230:18] it's like the Super Bowl running all the
[230:20] time compared to uh and take or sorry
[230:24] something running all the time taking
[230:25] that attention
[230:26] away from um you know real life creators
[230:31] and I don't have um
[230:35] I don't have good advice for how to
[230:38] safeguard ourselves from from AI and
[230:41] from generated content in any form
[230:44] whether it's uh just
[230:48] broad challenges of attention being
[230:51] sucked away or whether it's more
[230:53] specifically you know fem niche or tech
[230:56] niche or car niche or whatever niche it
[230:58] is.
[230:58] >> I think I think being explicit with your
[231:00] audience about what you're okay with and
[231:03] what you're not is genuinely helpful.
[231:05] Like you you've made it clear a number
[231:07] of times that you're actually not cool
[231:09] with your AI likeness being used. Um,
[231:16] your audience can be
[231:19] surprisingly helpful in these situations
[231:21] of just cuz you you can't be everywhere,
[231:24] but people that know what you're okay
[231:25] with
[231:27] can be many places. I don't know, weird
[231:29] sentence, but um, as long as those
[231:32] people can point it out, as long as
[231:33] those people can report it, it it it can
[231:36] be really helpful.
[231:38] I I've I've seen it before. So,
[231:42] yeah.
[231:43] >> Couple more for you. Hey, Lionus, are
[231:45] there ever any hardware I guess hardware
[231:49] stuff that you fear will be taken off
[231:51] the market or scalped because of a shout
[231:55] out you've made?
[231:56] >> Yeah, sure. I I actually checked this
[231:57] one ahead of time and I had a couple
[231:59] ready for you. I'm a big fan of the
[232:01] Rockat SOA lap board. It's, in my
[232:03] opinion, the best lap board.
[232:05] >> Oh, yeah. We have them all over the
[232:06] place here. Well, it's because I bought
[232:08] a couple of them because I knew that
[232:10] Rockat was going away and it was
[232:11] discontinued. Um, and it's These are
[232:15] both parts only for $55 and $85
[232:18] respectively. Uh, okay. I might have
[232:20] already done it then. Um, they used to
[232:22] be a little easier to get your hands on.
[232:23] I have like three of them just in case
[232:25] they die. And then another one is
[232:26] Optane. Dude, Optane is
[232:29] >> Yeah.
[232:30] >> is goated and especially feels
[232:33] especially relevant today. If only Intel
[232:36] and Micron could have held it together
[232:37] long enough to meet the demands of the
[232:39] AI boom, they would they would literally
[232:42] not be able to keep Optane on the shelf
[232:46] right now. the demand would be
[232:48] astronomical and it would also have been
[232:50] such a huge competitive advantage for
[232:52] their server hardware that would be
[232:54] compatible with like Optane DIMs and
[232:56] stuff like oh my god Optane was such a
[232:59] it was like the biggest missed
[233:00] opportunity of the last 10 years in in
[233:03] tech in my opinion. Okay, that's a
[233:06] that's a huge statement but it's it's up
[233:08] there. It's up there from like a
[233:10] financial standpoint. The missed
[233:12] opportunity here is like man, Intel just
[233:16] can't miss an opportunity to miss over
[233:18] the last little bit and Optane was by
[233:20] far the biggest one.
[233:24] And last one I got for you. Hey Lionus,
[233:27] what is your opinion on the new Final
[233:29] Fantasy Resonance that was recently
[233:31] announced?
[233:32] >> Um, my only opinion on it is that it's
[233:35] finally time once it comes out for me to
[233:38] play these damn games. I refuse to play
[233:40] part one of a series before the rest of
[233:43] it. I still have not read Game of
[233:45] Thrones because I am waiting for George
[233:48] RR Martin to finish writing these damn
[233:50] books, which it seems like he will never
[233:53] do. Uh I read the first one and then I
[233:56] was like, "Wait, he's not done writing
[233:58] them? That and I stopped reading them
[234:01] and um
[234:03] >> that's so funny.
[234:05] >> I'll wait. I I bought a box set of them
[234:08] and then realized he wasn't done writing
[234:10] it, so never opened it.
[234:13] >> Why are we so similar?
[234:15] >> I was I was hearing you describe this
[234:18] and then like the same series.
[234:22] Oh, that was pretty good.
[234:27] Oh man.
[234:29] >> For what it's worth, the first book's
[234:30] great.
[234:32] >> Yeah. No, I I'm looking forward to it.
[234:34] >> Yeah. Yeah,
[234:38] >> you guys. Um,
[234:41] that's all I
[234:42] >> I don't like getting into shows that
[234:43] aren't done either.
[234:44] >> I procrastinated through the entire Game
[234:46] of Thrones TV series
[234:48] >> and I felt so good that I was behind the
[234:51] entire curve.
[234:53] >> Yeah.
[234:54] >> Anyway, that's all I got.
[234:56] >> All right, that's all I got then. We
[234:58] will see you guys next week. Same bad
[235:01] time, same bad channel. Are we going to
[235:02] manage to both be here in studio next
[235:04] week? I think it's been almost a month.
[235:06] >> I'm there in studio. Are you there in
[235:08] studio?
[235:09] >> Nice.
[235:10] >> I think we're all here then. All right.
[235:11] Good stuff.
[235:12] >> Nice.
[235:12] >> See you guys there.
[235:13] >> Bye.
[235:30] Heat. Heat.