[0:00] On Wednesday, the White
[0:01] House announced its plan
[0:02] to make America
[0:03] the world leader
[0:04] in artificial intelligence.
[0:06] The main way it will seek to do
[0:08] that is to
[0:09] scale back on regulation of AI.
[0:12] I recently had a chance
[0:14] to sit down with Bill
[0:15] Gates to talk about
[0:16] how he sees
[0:17] the present and future of AI.
[0:21] Bill Gates,
[0:22] pleasure to have you on.
[0:23] We've been trying
[0:24] to track on the show
[0:25] the progress of AI
[0:27] and the place we are now at
[0:29] is everybody talks about
[0:31] AGI, general intelligence.
[0:34] What's the difference between
[0:36] artificial intelligence
[0:37] and what people are now
[0:38] predicting,
[0:39] which is
[0:40] artificial general
[0:41] intelligence?
[0:43] You know,
[0:43] the people use very different
[0:47] definitions and
[0:50] you can say when is AI
[0:53] so good
[0:54] that something like
[0:55] a telesales job
[0:57] or a support job that just
[0:59] I mean, the AI do that work,
[1:01] that it's
[1:02] way cheaper and more accurate
[1:04] than humans are.
[1:06] So that's,
[1:07] you know, really
[1:07] looking at the labor
[1:08] substitution piece.
[1:10] Or you can
[1:10] look at the most creative
[1:12] things that humans do,
[1:14] like come up with a new drug,
[1:17] that,
[1:18] helps with tuberculosis.
[1:20] Us.
[1:21] and,
[1:23] you know, look at okay.
[1:25] Is I just helping humans
[1:26] get this done,
[1:27] or is it eventually
[1:29] replacing humans?
[1:30] You know, like,
[1:31] people talk about writing code.
[1:33] Simple coding tasks.
[1:34] AI today can replace human work
[1:38] the most complex coding tasks.
[1:40] It's not able to do yet.
[1:42] And people
[1:43] in the field disagree
[1:46] is that.
[1:47] And within the
[1:47] next year or two,
[1:48] or is it more like
[1:49] ten years away.
[1:51] But I
[1:52] you know,
[1:52] I it's improving at a rate
[1:56] that surprises me.
[1:57] Things like this deep
[1:58] research capability.
[2:00] You know, a few times a day
[2:01] I take some complex question
[2:03] and just for fun, I see in the,
[2:07] the AI does
[2:08] an awfully good job.
[2:11] gathering
[2:12] all the materials and
[2:13] and bringing it,
[2:16] summarizing what
[2:17] I need to know.
[2:18] Satya Nadella says
[2:20] 30% of Microsoft's code
[2:21] is now being done by AI.
[2:24] That seems inevitably
[2:25] to mean one.
[2:27] You will hire fewer coders.
[2:29] You you'll need fewer of all.
[2:30] I mean,
[2:31] what is it
[2:32] that paralegals
[2:33] do at a law firm?
[2:34] They do discovery,
[2:35] which is essentially
[2:36] pattern recognition,
[2:37] is very easy for AI to do.
[2:39] Similarly, entry level
[2:40] people at accounting for right
[2:42] all that is going to mean
[2:44] white collar work.
[2:45] College educated, graduates
[2:48] are going to have
[2:49] a more challenging
[2:50] job environment.
[2:52] Well, when you
[2:52] when you improve productivity,
[2:54] there are you can make more
[2:58] and so it
[2:59] it shouldn't mean,
[3:01] you know,
[3:02] if you get less productive,
[3:03] that's bad.
[3:03] And if you get more productive,
[3:04] that's good.
[3:05] It means you can free up
[3:06] these people
[3:07] to have smaller class size
[3:09] or have longer vacations or,
[3:14] you know, to help do more.
[3:16] So it's it's not a bad thing.
[3:19] The question is, is it
[3:20] come so fast
[3:22] that you don't have time
[3:23] to adjust to it?
[3:25] And in parallel,
[3:26] you know, the blue collar work
[3:29] when the robotic arms
[3:31] start to be decent,
[3:32] which they're not today.
[3:34] And that starts to affect
[3:36] even larger classes of labor.
[3:39] So this is a
[3:41] a profound set of changes.
[3:44] you know, I'm
[3:45] working with Microsoft, OpenAI,
[3:48] on these things.
[3:50] You know,
[3:50] my lens is
[3:51] let's make sure it gets out
[3:53] in low income countries
[3:55] to help with their health
[3:56] and education and agriculture.
[3:58] So finally,
[3:59] what advice
[4:00] would you give to
[4:02] young people
[4:03] who are thinking about
[4:04] how do you enter
[4:05] the world of AI?
[4:06] What should the skills they
[4:08] they develop be?
[4:10] You know,
[4:11] if you get to very
[4:12] profound levels,
[4:14] you almost get
[4:15] to philosophical questions
[4:17] in this transition period.
[4:19] The ability to use
[4:22] these tools
[4:23] is both fun and empowering.
[4:26] You know, I, you state,
[4:27] have an advantage
[4:28] that I have very smart people
[4:29] I can call up
[4:30] when I get confused
[4:31] about physics,
[4:32] but now I actually use
[4:34] deep research.
[4:35] And then I'll send
[4:36] that answer to my,
[4:38] you know,
[4:39] smart friends and say,
[4:40] hey, did did it get it right?
[4:41] And most time they're like,
[4:42] oh yeah, you didn't need me.
[4:45] and so you can really learn
[4:47] so much.
[4:49] and, and the idea of the tutors
[4:51] that people like Khan
[4:52] Academy are building on
[4:54] how will get that out?
[4:56] all over the world.
[4:58] And so embracing this and,
[5:02] you know, tracking
[5:03] it will be very,
[5:04] very important.
[5:06] That doesn't guarantee that
[5:08] that we're not going to have
[5:09] a lot of, dislocation.
[5:12] But I really haven't changed
[5:13] my be curious, read
[5:14] and use the latest tools,
[5:16] recommendations
[5:17] for young people.
[5:18] Be curious, read
[5:20] and use the latest tools.
[5:21] In this case, AI. Absolutely.
[5:23] Bill Gates,
[5:24] pleasure to have you on.