7min chapter

Americano cover image

Is AI the new arms race?

Americano

CHAPTER

AI's Evolution and Economic Impact

This chapter explores the evolution of artificial intelligence, highlighting the release of ChatGPT and its effects on the tech landscape and investment sector. It delves into the competitive dynamics between major tech firms, economic ramifications, and differing perspectives on the future of AI technology and its reliance on computing resources.

00:00
Speaker 2
That appears to have calmed a little bit yesterday. And then today, we don't know. Am I right there?
Speaker 1
Yeah, I mean, I think that's right. It's just a tremendous, tremendous story for so many levels. I mean, look, you know, starting about late 2022, so a little over two years ago, OpenAI released ChatGPT, which was sort of the first major light bulb moment for people like, wow, this is incredible technology. You ask a question into a box, it basically does whatever you want to varying degrees of quality. That opened up this incredible sort of multi-year rally in all kinds of companies that were in some way connected to artificial intelligence. Still very unclear how anyone is going to make money on this. Still anyone clear where any of it is going. Nonetheless, it's just so exciting, so novel to people that all the big tech companies, you know, who are building models, they've rallied. NVIDIA has become the biggest company in the world because they make the very best chips for training these models and so forth. And it's gotten so big that, you know, there is a sense in which the entire U.S. stock market, to some extent, is like a bet on AI because these companies have just become so big. And we know that Chinese companies of various sorts are extremely competitive and eating America's lunch and eating the rest of the world's lunch and things like cars and solar panels and things like that. We've known that for a long time. But the truth of these matter is that cars and solar panels just aren't that important to the U.S. stock market. These companies aren't that big. They're not that profitable and so forth. And then DeepSeek is released. Now, weirdly, and we can talk about this too, the actual performance of the paper and its cheapness and so forth were actually spelled out in December, but maybe because it was Christmas and maybe because there wasn't a public chatbot, it sort of didn't get much attention. Suddenly, people start paying attention to the chatbot. It's very cheap. For almost any purpose, it seems to be just as good as not the most advanced model from OpenAI, but one of the more advanced models. And suddenly, it's like, wait a second, we're spending hundreds of billions of dollars on all this stuff. Can it be made much more cheaply? And so you get this big stock market sell-off on Monday, which strikes all kinds of nerves from just pure finance to economics to national security. It's a great story. Well,
Speaker 2
it is a great story because it does show just how manic things are around AI and how much money is going into how so little understanding. Because as you say, DeepSeek does seem to have been around for a while and then suddenly there's panic because it goes a bit public. And NVIDIA's role is, there's a lot of talk about that too, isn't there? Because some people are saying, I believe, correct me if I'm wrong, that the arrival of DeepSeek should be good news for NVIDIA. That's right. And yet it was seen as very bad news for NVIDIA. Can you explain that?
Speaker 1
Absolutely. So I think there's like the NVIDIA question, as it were, is sort of multiple dimensions. So one is we don't really know for sure the hardware capacity that DeepSeq currently has access to. Now, something that's very interesting about DeepSeq is that it was born out of a quantitative hedge fund in China. And there is this very elegant story that people like to tell. And it may even be true that, you know, speculative finance in China isn't as encouraged as it is in the West, and maybe they felt some pressure to do something productive with all their brains and all the chips that they have access. It's a nice story. I kind of want to believe it. I kind of think it's one of these stories that's a little bit too nice to, you know, satisfies our bias about this sort of, you know, this sort of cost of financial capitalism and so forth. But it's a nice story. But it is true. There was this quantitative hedge fund, and they seem to, on the side, have developed a large language model. They probably had access to some volume of NVIDIA chips, in part because there weren't really any export restrictions on some of the older models up until recently. I guess there are two questions, and there are sort of two ways to think about it. If you think that AI is just, say, some business tool that will be used to, say, for example, reduce the need for humans to do tech support at large organizations, and that there are some tools that exist and AI can solve that, then you would say, well, that's bearish for NVIDIA because it turns out that if we believe all these stories, you just don't need as much chips and computing capacity to do some of these as we thought. On the other hand, if you say, no, ultimately, this is going to head towards something towards AGI, which I don't really know what that means, but these supercomputers that's sort of like smarter, 10 times smarter than a human, and we're still on the race towards that. Then the other story is that you can do a lot with NVIDIA chips, and we just are going to need even more of them because we're going to use the algorithmic techniques that DeepSeq discovered to get even more out of these chips. But in the end, we're just going to use up this compute. And so in the end, we'll just be using all the same NVIDIA chips. We'll just be getting even more out of them. I don't know which story is true. I don't know. You know, I sort of think it's almost like there's two. I get the impression in a sense that there is this sort of like Silicon Valley mindset and the Wall Street mindset. And the Wall Street mindset is like, OK, what is the business case for this technology? Can this be used to save money on human labor and improve productivity and improve margins, right? And then the Silicon Valley story is, no, you don't understand. We're building the new God here, and we're not going to rest because we're going to save some doctor's office some money because they didn't have to have a person booking the appointments. Yeah, maybe we could do that. But our goal here is to build the new God, and none of this slows down anything. In fact, it just accelerates it. And I don't know how much investors really believe the Silicon Valley story. But, you know, there are different stories you can tell about what we're actually doing here with all this technology. But important to mention here is that everybody is still, except NVIDIA and a few other like infrastructure companies, are basically still losing money on all of this tech. But so far, nobody has really seemed to care.
Speaker 2
Yes.

Get the Snipd
podcast app

Unlock the knowledge in podcasts with the podcast player of the future.
App store bannerPlay store banner

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode