0:00And I think AI will probably increase the global economy by 20 to 30%. That's my rough rough estimate. Um meaning on the order of 20 to 30 trillion dollars per year. Um and and AI will be able to do anything digital. Anything that does not require shaping of atoms by hand. Um probably by the end of next year. So as as I'm sure people know uh AI is already uh incredibly good at software.
0:33Um and it it's getting to the point where AI won't just be good at software. It'll be what I call Stockfish-level good. So Stockfish is a chess program that can beat the world's best chess players very easily. In fact at this point you could run Stockfish on your phone and beat Magnus Carlsen at chess.
0:53So uh at some at some point next year is my prediction uh software will be so good AI software will be so good that it will be Stockfish-level good. Meaning that it is impossible for a human to compete um in writing software uh with AI. That like AI will just crush all humans at software. And I think it will it will be uh extremely good possibly Stockfish-level but but but certainly extremely good at all forms of of engineering and anything
1:22digital. Uh literally in 12 12 to 18 months.
1:28I one of the stats So so so Actually, I'll let me I just want one more thing to add. My apologies. Um the the So 20 20 to 30% increase of total global economy. So this is quite a lot of prosperity we're talking about here. Um uh just just from digital AI. But from robotics uh from um from humanoid basically think of it like a general-purpose robot robot AI.
1:53I think we'll see many multiples of the global economy. Meaning like you you can increase the economy by a factor of 10 or more. The these are mind-boggling numbers.
2:05Yeah, I was just about to say that I mean one of the stats we gave to the group here was that in the four years since sort of ChatGPT was launched, I think we there's you know, over a billion people around the world already using AI. There's been very quick uptake around around the country on around the world on that. I guess the question to you is I think a lot of people after being in this sort of applied or or sort of physical AI, where where do you see sort of robotics trending? How quickly is it going to be implemented? I think I remember in the first Trump administration, we were talking about
2:34sort of automated factories and now we're a year later and 10 years later.
2:38So, how how how you know, how quickly do you think these these changes are happening in the sort of physical AI world?
2:44Yeah, so so anything physical always takes longer than anything which is digital. Digital, you know, it's it's when when when you sell something digitally, it's the software that you can easily copy across to other other computers. Um when it's physical, you've got to build up an entire massive supply chain. You've got to you've got to move a lot of a lot of atoms.
3:02Um and and so supply chains are very much global at this point. So, uh that's why it takes longer.
3:08Nonetheless, um when you think of humanoid robotics, the way you think about it, I think I I think the right framework is consider that the usefulness of a humanoid robot, a general purpose robot, is going to be um roughly the digital AI software how how how good is the AI software times how good is the AI chip in the robot times how good is the electromechanical dexterity, especially of the hands. Um now, all three of those
3:37things are improving exponentially. Um and and the usefulness of the robot is those those three things multiplied by each other. Then, when you make the robots, the robots will um um Well, robots will start manufacturing the so you get a recursive effect. So, it it starts out very very slowly, but then it grows um at an explosive rate.
4:02Um so, if you you say like 10 years from now, I would say there are well over a billion humanoid robots.
4:10Um and the productivity per robot will be probably five times that of a human. Meaning meaning that the productivity in 10 years of humanoid robots, and I think this is a conservative estimate, by the way. Um This is this is what I I I'd be willing to be willing to put to put serious money betting on. Um that there will be at least a billion robots in 10 years.
4:35And that those robots will be at least five times the output of a human. Meaning the the the the billion humanoid robots will will um be more productive than all humans combined.
4:47Wow. Um shif- shifting gears just a little bit to to to questions kind of facing the US today. Um you know, data centers have been uh a big political issue over the over the last um 6 to 8 months here in the United States, and I think there's a a general understanding that um in order to to drive and power the the the AI revolution that's coming, we need to have the the electricity um and the data center uh compute capacity um to to do the training and the inference of all
5:17this AI. Um you know, how how do how do you think about this particular issue?
5:21Um where where are we on the on the curve of of kind of how much we built versus how much we need? And as government leaders here think about how to prepare their economies and and build the right power and data infrastructure, how should they be thinking about where sort of compute and and power needs are going to be in the in the next few years?
5:40Well, there actually is a quite a crisis of power. So, the This is a And in fact, this is something if you if you um just if you just sort of follow the AI topic on the X platform, which by the way is is where almost all of the AI discourse takes place. Um You I think you get a very good sense for where things are headed.
6:04Um That's how that's how I get my news and it's it's incredibly good. Everyone who's anyone in AI posts on X. Um so so so that's why I'd recommend like just just just go on the AI AI AI topic on X and you'll you'll understand all these these things and get a day day by day account of things.
6:23The the uh And and and what the consensus is at this point is that the there will be a significant power shortfall next year. So not like distant future. Uh there's expected to be I believe I believe the the consensus estimate among analysts that follows the AI space very closely is that there will be at least a a 15 gigawatt uh shortfall of power uh in 2027 for AI chips.
6:51So this is perhaps um an obvious uh thing that one would expect to occur because the rate at which AI chips is being produced is is has been rising incredibly rapidly. They're sort of rising on the order of 40 40 to 50% a year. But the but the power available outside of China has been rising at like 10 to 20% a year.
7:15So obviously uh the the faster rising thing will will eventually overwhelm the slower rising thing. And so um in fact, even I'd say at this point there there there are challenges with power even before next year, which is why uh Google and Anthropic and many other companies are actually leasing compute from SpaceX. Um because we've been able to turn on um AI better than anyone else so far.
7:42And that this is by by constructing our own power plants. Is the only way we were able to do it.
7:47Um so now now China does have a tremendous amount of electricity, but but but uh due to GPU export bans, one cannot uh you know, establish data centers with the latest chips in China. So so I I but so really the consideration is what what sort of electricity growth is there outside of China?
8:08Um and that is is currently a significant shortfall relative to AI chip production.
8:14So there is this this creates an opportunity, I think, for countries around the world just to to say if to to to sort of if they're interested in AI data centers, um to to to construct uh a lot of power um and uh and offer that to uh AI companies. Um and in exchange, of course, that these AI data centers would be would be taxed and and have to pay, you know, reasonable fees and stuff.
8:40Uh but it does create an opportunity for for a lot of countries.