The AI Future No One Wants to Talk About

[00:03] So you think that once artificial super 

[00:09] a month and henceforth live super intelligently 

[00:15] And today I want to tell you why. The recent 

[00:22] former chief AI scientist at Meta, quit because 

[00:29] now founded his own company. The US government 

[00:36] model, Fable, for all but US citizens, giving 

[00:42] hoped for, entirely for free. And Grok supposedly 

[00:49] Iran. I think all these developments tell us 

[00:54] happens in artificial intelligence is basically 

[01:00] pressure. The software and hardware of machine 

[01:05] become extremely expensive to use and maintain and 

[01:11] already see the beginning of this. The future is 

[01:18] is your genius assistant is currently unavailable 

[01:24] avoidance. The artificially intelligent systems 

[01:30] useful in some domains like coding and textbased 

[01:36] general intelligence, but we'll get there. Maybe 

[01:42] in a decade we'll have machines that are much 

[01:48] the machines become more intelligent or we become 

[01:54] risks posed by artificial super intelligence 

[02:00] Indeed, they almost certainly will and then they 

[02:06] which won't be all that difficult because it's 

[02:11] without any intelligence at all. But this is not 

[02:16] I want to talk about what super intelligence will 

[02:24] we'll have to deal with first. Think back 250 

[02:31] didn't just get richer. They held power over the 

[02:37] which basically started the wave of global 

[02:43] They didn't invent colonalization, but they 

[02:48] Germany followed their example decades later and 

[02:55] is about to happen with AI, not with ships and 

[03:02] nations, or corporations to develop super 

[03:07] economically, militarily, and intellectually. 

[03:13] understands it. The rest of the world clearly 

[03:19] idiotic things like building bigger particle 

[03:23] you know I'm me. More seriously, at this point in 

[03:30] matters. Whoever gets there first will literally 

[03:39] we currently use, it's tempting to extrapolate 

[03:46] access to the newest models eventually. Yes, 

[03:51] reasons. Okay, but look, unless you want to breed 

[03:57] affect you. And if you do, well, you might 

[04:03] companies have strong commercial incentives to 

[04:10] is a temporary phase that won't continue. The 

[04:17] phases. The training that takes a long time and is 

[04:24] in the weights that can be copied easily be widely 

[04:30] models are nowhere near being profitable, but 

[04:36] useful enough, they'll get there eventually. And 

[04:43] dramatically change. To see what's likely coming 

[04:50] the training of these models takes a lot of 

[04:56] meaning a lot of hardware in the form of chips 

[05:03] requirements are already a bottleneck that's been 

[05:09] have serious shortcomings. Most importantly, 

[05:15] then you roll out the update, then you train a new 

[05:22] these models with tools, giving them memory and 

[05:29] have a basically unfixable safety problem from 

[05:36] we'll see a switch to an entirely new basic 

[05:43] and Google Deep Mind and Yann LeCun's new company 

[05:50] in which the artificial mind basically learns 

[05:55] did during evolution. That, so the hope, will teach 

[06:03] ultimately be the step to general intelligence. 

[06:09] they suffer from what's been called catastrophic 

[06:16] knowledge. They overwrite it. The human brain has 

[06:22] parts. In my case, one part does physics, one part 

[06:28] cheese. And indeed the trend in large language 

[06:34] to using task dedicated subsections of the 

[06:41] more and more properties from the human brain 

[06:48] the trend is now towards continuous learning. 

[06:54] in hardware already. Google in particular has 

[07:00] and multiple companies are working on what's been 

[07:07] purpose with the hardware design. The major reason 

[07:13] Now let's put these things together. The economic 

[07:19] that it becomes less energy intensive and will 

[07:26] that we already see tells us that this will mean 

[07:32] interwoven not unlike in the human brain. The 

[07:39] then becomes quite similar to the metabolic 

[07:45] The thing is now though that the more the 

[07:51] with its hardware and the bigger it becomes, 

[07:57] that this will become impossible, just difficult 

[08:03] except for maybe the occasional backup. Where does 

[08:10] of this development is a megabrain architecture. 

[08:18] central model. And from this megabrain, 

[08:24] models for routine use to be deployed elsewhere 

[08:31] ask you to rate your experience. But don't worry, 

[08:37] a sale on our store with 25% off on pretty much 

[08:45] the elements of the human body. Megabrains are 

[08:52] think that's not a coincidence. It's what you get 

[08:57] will be subject to similar environmental 

[09:03] The real question is then what's the economics 

[09:10] super brains will be extraordinarily expensive. 

[09:16] few companies and maybe a few governments who will 

[09:22] restricted. Not only will you need some safety 

[09:28] hand over a lot of money. It's a future in which 

[09:33] afford it. A bold new concept, also known as the 

[09:40] be that the rich will get richer and the poor 

[09:46] means that we'll increasingly live in a world in 

[09:53] how or why. new materials, new technologies, new 

[10:00] them that we can't understand with decisions being 

[10:07] remain in charge. There'll be some fractions 

[10:13] instead insist on living in low tech AI free 

[10:19] the future will likely bring. artificial super 

[10:26] many. And we can finally work on our ultimate 

[10:32] issue with artificial intelligence is the rapid 

[10:39] today's sponsor, Ground News, can help you 

[10:44] that collects news from all over the world. It 

[10:49] have to sort through a dozen headlines on the 

[10:55] fact check for all the coverage. And ground news 

[11:01] glance. An interesting recent example is this 

[11:07] back several more offshore wind leases to instead 

[11:13] a quick summary of all the coverage here and 

[11:18] covered on the political right. Ground News also 

[11:24] tells you who owns the media outlets and tells 

[11:30] has this interesting feature called blind spot. 

[11:36] on one side of the political spectrum. I found 

[11:43] context. And of course, I have a special offer for 

[11:49] which gives you access to all their features. All 

[11:56] or use the QR code so they'll know I sent