Bill Gates on navigating an AI future

[00:00] On Wednesday, the White

[00:01] House announced its plan

[00:02] to make America

[00:03] the world leader

[00:04] in artificial intelligence.

[00:06] The main way it will seek to do

[00:08] that is to

[00:09] scale back on regulation of AI.

[00:12] I recently had a chance

[00:14] to sit down with Bill

[00:15] Gates to talk about

[00:16] how he sees

[00:17] the present and future of AI.

[00:21] Bill Gates,

[00:22] pleasure to have you on.

[00:23] We've been trying

[00:24] to track on the show

[00:25] the progress of AI

[00:27] and the place we are now at

[00:29] is everybody talks about

[00:31] AGI, general intelligence.

[00:34] What's the difference between

[00:36] artificial intelligence

[00:37] and what people are now

[00:38] predicting,

[00:39] which is

[00:40] artificial general

[00:41] intelligence?

[00:43] You know,

[00:43] the people use very different

[00:47] definitions and

[00:50] you can say when is AI

[00:53] so good

[00:54] that something like

[00:55] a telesales job

[00:57] or a support job that just

[00:59] I mean, the AI do that work,

[01:01] that it's

[01:02] way cheaper and more accurate

[01:04] than humans are.

[01:06] So that's,

[01:07] you know, really

[01:07] looking at the labor

[01:08] substitution piece.

[01:10] Or you can

[01:10] look at the most creative

[01:12] things that humans do,

[01:14] like come up with a new drug,

[01:17] that,

[01:18] helps with tuberculosis.

[01:20] Us.

[01:21] and,

[01:23] you know, look at okay.

[01:25] Is I just helping humans

[01:26] get this done,

[01:27] or is it eventually

[01:29] replacing humans?

[01:30] You know, like,

[01:31] people talk about writing code.

[01:33] Simple coding tasks.

[01:34] AI today can replace human work

[01:38] the most complex coding tasks.

[01:40] It's not able to do yet.

[01:42] And people

[01:43] in the field disagree

[01:46] is that.

[01:47] And within the

[01:47] next year or two,

[01:48] or is it more like

[01:49] ten years away.

[01:51] But I

[01:52] you know,

[01:52] I it's improving at a rate

[01:56] that surprises me.

[01:57] Things like this deep

[01:58] research capability.

[02:00] You know, a few times a day

[02:01] I take some complex question

[02:03] and just for fun, I see in the,

[02:07] the AI does

[02:08] an awfully good job.

[02:11] gathering

[02:12] all the materials and

[02:13] and bringing it,

[02:16] summarizing what

[02:17] I need to know.

[02:18] Satya Nadella says

[02:20] 30% of Microsoft's code

[02:21] is now being done by AI.

[02:24] That seems inevitably

[02:25] to mean one.

[02:27] You will hire fewer coders.

[02:29] You you'll need fewer of all.

[02:30] I mean,

[02:31] what is it

[02:32] that paralegals

[02:33] do at a law firm?

[02:34] They do discovery,

[02:35] which is essentially

[02:36] pattern recognition,

[02:37] is very easy for AI to do.

[02:39] Similarly, entry level

[02:40] people at accounting for right

[02:42] all that is going to mean

[02:44] white collar work.

[02:45] College educated, graduates

[02:48] are going to have

[02:49] a more challenging

[02:50] job environment.

[02:52] Well, when you

[02:52] when you improve productivity,

[02:54] there are you can make more

[02:58] and so it

[02:59] it shouldn't mean,

[03:01] you know,

[03:02] if you get less productive,

[03:03] that's bad.

[03:03] And if you get more productive,

[03:04] that's good.

[03:05] It means you can free up

[03:06] these people

[03:07] to have smaller class size

[03:09] or have longer vacations or,

[03:14] you know, to help do more.

[03:16] So it's it's not a bad thing.

[03:19] The question is, is it

[03:20] come so fast

[03:22] that you don't have time

[03:23] to adjust to it?

[03:25] And in parallel,

[03:26] you know, the blue collar work

[03:29] when the robotic arms

[03:31] start to be decent,

[03:32] which they're not today.

[03:34] And that starts to affect

[03:36] even larger classes of labor.

[03:39] So this is a

[03:41] a profound set of changes.

[03:44] you know, I'm

[03:45] working with Microsoft, OpenAI,

[03:48] on these things.

[03:50] You know,

[03:50] my lens is

[03:51] let's make sure it gets out

[03:53] in low income countries

[03:55] to help with their health

[03:56] and education and agriculture.

[03:58] So finally,

[03:59] what advice

[04:00] would you give to

[04:02] young people

[04:03] who are thinking about

[04:04] how do you enter

[04:05] the world of AI?

[04:06] What should the skills they

[04:08] they develop be?

[04:10] You know,

[04:11] if you get to very

[04:12] profound levels,

[04:14] you almost get

[04:15] to philosophical questions

[04:17] in this transition period.

[04:19] The ability to use

[04:22] these tools

[04:23] is both fun and empowering.

[04:26] You know, I, you state,

[04:27] have an advantage

[04:28] that I have very smart people

[04:29] I can call up

[04:30] when I get confused

[04:31] about physics,

[04:32] but now I actually use

[04:34] deep research.

[04:35] And then I'll send

[04:36] that answer to my,

[04:38] you know,

[04:39] smart friends and say,

[04:40] hey, did did it get it right?

[04:41] And most time they're like,

[04:42] oh yeah, you didn't need me.

[04:45] and so you can really learn

[04:47] so much.

[04:49] and, and the idea of the tutors

[04:51] that people like Khan

[04:52] Academy are building on

[04:54] how will get that out?

[04:56] all over the world.

[04:58] And so embracing this and,

[05:02] you know, tracking

[05:03] it will be very,

[05:04] very important.

[05:06] That doesn't guarantee that

[05:08] that we're not going to have

[05:09] a lot of, dislocation.

[05:12] But I really haven't changed

[05:13] my be curious, read

[05:14] and use the latest tools,

[05:16] recommendations

[05:17] for young people.

[05:18] Be curious, read

[05:20] and use the latest tools.

[05:21] In this case, AI. Absolutely.

[05:23] Bill Gates,

[05:24] pleasure to have you on.