00:00People who have concerns, you know, I think those are warranted as well because there are going to be risks
00:07that need to be managed along the way.
00:09And I think that needs to be done very thoughtfully.
00:12But fundamentally, I mean, I think that there are a few big questions.
00:16One is, do you kind of, are we optimistic about having more capability, more distributed to a lot of people?
00:25And will that, do we think that that's going to make it so that people can improve their own lives
00:33and improve their own economic welfare?
00:35And then the second that I think is a big question is basically just this question of invention, which is
00:44how much more do we think we have to invent and do individuals have to invent?
00:49And I think overall, if you have this view that we're pretty early in this, right?
00:56I think most people have things that they want to see exist in the world, but that it's pretty hard
01:01to create today.
01:03And I think AI and the agents that people have will basically just make it a lot easier for people
01:11to accomplish all of these goals and create the different things that they want in their lives,
01:16whether that's kind of improving things in their personal lives or creating new businesses, making scientific discoveries, learning new things
01:23when you're a student.
01:25I mean, all of that is just going to be a lot more possible.
01:28And I think if you have this view that there's way more to invent ahead of us, then you should
01:34be very optimistic that AI is going to help us do that.
01:39The alternative, of course, is that if AI can help us invent things and instead is primarily a technology that's
01:47used to automate the current things that people can do without creating new things,
01:51then that has a lot of problems that I think have been discussed widely, but I think as long as
01:57you kind of have this belief that I do,
01:59that there are a nearly infinite number of things that people want to invent and create in their lives to
02:05improve their lives and drive...
02:16...then everyone will have the tools to create new things beyond what's possible today and the economy will be more
02:23dynamic and generate more broad-based prosperity.
02:26And I think that's kind of what you're talking about here.
02:28I mean, can you expand on that a little bit?
02:30I think many times people view, as you said, AI as something that is just automating existing tasks versus, you
02:36know, actually inventing new things and bringing much broader, broader-based prosperity.
02:41Yeah.
02:42And again, my philosophy is basically three things.
02:46One is that empowering individuals is the source of prosperity.
02:51Two is that the purpose of AI and superintelligence is invention, not automation.
02:57And three is that the foundation for safety in the future is achieving the right balance of power and having
03:05the technology diffused such that you can set up the system of checks and balances rather than restricting it to
03:12only being available to a small number of people or companies or governments.
03:16So, in terms of the prosperity piece, I mean, I basically think that if you ask, you know, most people
03:26have things that they want to create in the world.
03:28And the reality is that today it's just hard.
03:31I mean, we have a lot of good technology, but using the technology is also hard, right?
03:35I mean, it's powerful, but you have to go learn how to use all these different tools.
03:39And one of the things that AI can do is it just stitches these pieces together.
03:44And we're working on agents at Meta that we want to put in the hands of billions of people where
03:50they just understand your goals and can work 24-7 on your behalf to help you achieve them,
03:56whether that's improving aspects of your personal life or functioning almost as a sort of business in a box where
04:05you give it an idea, you steer it on the creative direction of what you want to see,
04:09but it takes care of a lot of the kind of basic technical aspects of kind of running a business.
04:17And I think that this is just going to be a very powerful thing.
04:20So I would expect that the cost of starting a business is going to go way down.
04:24It's going to take way less capital in the future, maybe way less people to get started.
04:30And therefore, the shape of the economy going forward, I think that there are going to be a larger number
04:36of startups and a larger number of thriving small businesses.
04:42And I think that that's a pretty good foundation for the economy.
04:45So that's one that I'm pretty optimistic about.
04:48Now, I do think we're going to go through a period of change.
04:51I mean, the way that large companies work is going to evolve.
04:55Not every part of that is going to be positive.
04:58We're going to have to make sure that we invest in making sure that people can use all these new
05:02tools, and that's something that we care a lot about.
05:05But I do think that the ability to start new things is going to be immense, and that's going to
05:09lead to this very dynamic period.
05:11So I think it leads to wanting to have policies that make it easy to create things, easy to evolve
05:18and change the systems that we have today.
05:22But, yeah, I'm very optimistic about that.
05:24There's the other part of this, which is in terms of skilled trades to build all the data centers that
05:29we need.
05:29And that's going to be hundreds of thousands and maybe millions of jobs just because the demand for data centers
05:39is immense.
05:41I mean, what we've basically found at Meta is that we can't find the volume of skilled tradespeople that we
05:50need to build the data centers, even just that we as one company are trying to build.
05:53So we created this program, America's Workforce Academy, that basically helps train people in skilled trades.
06:02I mean, we need way more carpenters, way more electricians, people who are skilled to build these complicated buildings.
06:08And so we offer this training program, and we guarantee you a job building one of our data centers at
06:13the end if you get through it.
06:16And it's been a big success so far.
06:18So I think that things like that are going to be part of how we create this upskilling in the
06:24economy.
06:25But overall, my guess is that this is going to create way more jobs.
06:28I just think it's going to be a very dynamic period.
06:31Yeah, I think your point about workforce resonates with a lot of folks here at this table.
06:37Because I think sometimes when we think about the AI revolution, we're not thinking about the welders and the fiber
06:42technicians and the construction workers that all go into building infrastructure that supports it.
06:47And I know the program that you guys launched, America's Workforce Academy, I think shows a different and important dynamic
06:53about how the private sector can be integrally involved in the training and skilling of individual Americans.
06:59And it's something that I think if you roll back, you know, two or three decades was something that was
07:03not even in the card.
07:04So it's incredible work.
07:06I think one thing that I wanted to kind of close with, which is interesting to a lot of folks
07:11here, is this question of open source.
07:15And, you know, you've announced your plan to open source and use 1.2, which is the leading sort of
07:21American open source LLM at the moment.
07:24So are you committed to doing more than this in the future?
07:26Or how do you kind of think about the role that open source will play in our ecosystem?
07:32Yeah, I mean, to me, open source is one important part of making sure that the AI future and the
07:38prosperity is broadly shared by a lot of people.
07:41So probably the biggest part of what we need to do is build these kind of agents that I talked
07:45about and make sure that they're distributed to billions of people.
07:49So that way the power and the capability of AI should be something that everyone has, not just something that's
07:57restricted to a small number of people.
07:59But the ethos of open source is also very aligned with this philosophy and it's something that Meta has focused
08:07on for a long time as well.
08:08It basically puts the power in individual developers' hands so that way they're not just dependent on a small number
08:15of these frontier labs for what is an increasingly critical part of the applications and technology that they're building.
08:24So I think that that's a very important part of what we do.
08:27So at Meta, you know, we build both open and closed models.
08:32We've had a long history of building open source software.
08:35We felt that it's worked well.
08:36When you look at issues like cybersecurity, I think it's just very well established over decades that open source systems
08:43end up being more secure because there's more scrutiny on them, right?
08:47It's basically if more people can see how the thing works and can measure it and understand it, understand where
08:53it works well, understand where the defects are, patch it, then that leads to systems that are more hardened over
09:00time.
09:00And I think that that's going to be the case here as well.
09:02So I think that this is an important part of having a secure and safe future is both having the
09:08checks and balances that come from having the ecosystem that isn't just centered around a small number of frontier labs,
09:14but that where, you know, everyone from governments to startups to individuals can access models that they control.
09:21And having that system kind of having the underpinning of it be something that everyone can scrutinize, I think, is
09:26going to be more likely to lead to a safe future as well.
09:28So we will continue doing some open source models.
09:32We kind of evaluate it on a model-by-model basis to understand whether the capabilities of what we're building
09:38are appropriate to open source at that point.
09:40But our philosophy in general is that we think that this is a really important ecosystem to make sure that
09:47we foster.
09:48And, you know, beyond just what we release ourselves, we're very strong advocates that open source shouldn't be something that
09:55people try to restrict or lock down.
09:57I think it's a very important part of creating a good future for everyone.
10:00Amazing. Well, thank you so much for your time, and I encourage all the ministers here to certainly look into
10:04America's Workforce Academy.
10:06I think it's one of the shining examples in the U.S. of how we can come together to train
10:11our workers for the jobs of the future.
10:13So thank you so much.
