You Ask, I Answer: Is There An AI Bubble?

In today's episode, let's answer Bob's question. Bob asks, is there an AI bubble? Well, that's a good question. And the answer is it depends on how you define a bubble. Bubbles in the investment are generally governed by five things, right?

Rapid price increases, meaning there's a lot of money flowing in. Uh speculative trading, where people are buying on what they think the future value of something is rather than the value it has today. Lots of people rushing in. So high volume is number three. Number four is decoupling from fundamentals, where the value of something is perceived as way higher than it's actually worth.

And number five is a uh this time it's different mentality. Bubbles are nothing new. I mean, bubbles have been around for centuries. Remember uh Holland's tulip bubble in the 1600s, I think it was 1600s, where tulip bulbs were just traded at ridiculous prices. Some variants of tools were traded at like a year's worth of income for a single tulip bulb.

That clearly fits the speculative trading part, the rapid price increases, high trading volume, and especially decoupling from fundamentals. Like a tulip bulb, no matter how nice that tulip is, tools aren't really useful other than decorative. I mean, you can technically eat them, but they're not very good. Um, definitely decoupled from fundamentals. So, with that lens, is AI a bubble right now?

Yeah, I mean, you'd kind of have to be blind not to see it, metaphorically speaking. Um the stupid money has arrived. And by that I mean seeing rapid valuation increases, seeing people making speculative investments, speculative trades, uh, a lot of people rushing into the space. All those folks who were, you know, crypto and NFT experts two years ago are suddenly AI experts today and uh are hawking AI services. Um the big one is decoupling from fundamentals, where what people are spending money on and uh and how they're investing is not based in reality on the valuation of a company.

Companies like Anthropic and OpenAI are wildly unprofitable, like wildly unprofitable. They are burning money like crazy, and the only reason they're still in business is because investors keep pouring money into them in the hopes that um someday they will be profitable. But there really is that huge decoupling from fundamentals. And especially with AI, there is the excessive optimism and the this time it's different uh mentality. In some ways, I can agree with that.

AI is a very different technology and knowledge work than we've ever seen before. It is very similar to automation in agriculture. Uh, up until we had the steam engine and things like the cotton gin stuff, humans were doing all of the working in agriculture. And then we have more and more machinery to the point where today one human being can do the work of 10,000 people, right? If it takes a uh it was one person can make can harvest a bushel of corn in about 10 hours.

One person driving the John Deere X91100 combine can harvest seventy two hundred uh bushels an hour or seventy-two thousand bushels in 10 hours. Right. So that is a 72,000% increase in productivity for that one person. And as a result, there are far, far fewer people who work in agriculture these days than they used to be because the machine machinery's gotten so efficient. AI is kind of the same thing.

Agriculture is a really good example because you have certain tools, right? A shovel makes things easier. But at a certain point, they crossed over into the tool doing the work for the person, right? You get a shovel or a rake, no matter how nice it is, that is you are still doing the work. You get into the John Deere X91100.

Yeah, you're steering it, but it's doing the work, right? It is picking the corn and harvesting it, cutting it and tossing it in a in a big old bin. And so it's doing the work. And that is the big flip, the big change there from human to machine. Just like robots on an assembly line.

Assembly lines used to be all human-powered, right? People just riveting the same five bolts every single day, you know, every hour every day. And robots now do the work. Humans supervise the work, but manufacturing and automation, the humans don't do the work anymore in some instances. The machines do the work and the humans supervise machines.

Knowledge work is experiencing the same change now, where you write a blog post or write an accounting report or balance the books or write a contract. Well, now a machine is doing the work and you are supervising the machine. And the same thing that is held true from time immemorial of humans doing the work to humans supervising something doing the work is going to hold true in AI. So from that perspective, it is different. But there is still definitely excessive optimism.

There's definitely a whole lot of people saying a whole bunch of things that are totally unfounded in many ways. It is possible, as Sam Altman from OpenAI talked about, that 95% of knowledge work jobs could be done by machine. In the same way that it is possible that, you know, and it did happen that 95% of agriculture jobs are now being done by machine in many different high yield scenarios. Big fields of wheat and corn and stuff. That's all harvested by machine now.

And yeah, you still need people to run the machines and to supervise to inspect, but you don't have somebody out there like yanking ears of corn off the plant anymore. So is there an AI bubble? Yes, there's absolutely an AI bubble today. When's it gonna burst? When these companies run out of money, when when they continue to operate in such a way that they are deeply unprofitable and basic economics takes over, there is only so many so long that a company can run at a loss before it's done.

Um and this will this is partly driven by things like you know excess money floating around, still left over from the pandemic and stuff. But there is absolutely an AI bubble, it will burst at some point, and what you should do about it is be aware of how the different technologies that are at play and make sure that you got backups. If there's a vendor in particular that you depend on, make sure you got a backup so that if that vendor goes away, you're not left holding the bag. That's gonna do it for this episode. Thanks for tuning in.

We'll talk to you soon. If you enjoyed this video, please hit the like button. Subscribe to my channel if you haven't already, and if you want to know when new videos are available, hit the bell button to be notified as soon as new content is live.