Transcription
[Music] Al Jazer podcasts.
Today, the artificial intelligence boom is showing cracks. Investors are weighing market valuation concerns and growing volatility in AI stocks. Top bank CEOs also warning of a potential market correction.
>> So, is the bubble about to burst or will AI live up to the hype? I'm Manuel Rapalo and this is the take.
[Music]
Hi, before we continue today's show, remember to leave us a comment telling us what you think about this episode and what stories you want us to do next. Or on your podcast app, leave us a review telling us where you're listening from. And give us a five-star rating while you're there. It helps the show.
>> My name is Paul Ford and I am the president of a board, an AI startup that builds software with AI. Paul Ford, welcome to the take.
Artificial intelligence seems to be everywhere right now. It's in research labs. It's in classrooms. It's in offices. Obviously, it's all over social media. We're seeing AI videos at this point and AI voices that are so realistic, it's hard, if not impossible to tell the difference between what's AI and what's real life anymore.
But at the same time, what also might not be so real is the hype. It's this huge growth in the value of AI companies. We're seeing headlines warning of an AI bubble. Even Sam Alman, who is the CEO of Open AI, the parent company of Chat GPT and of Sora, told The Verge that AI is in a bubble. What does that mean?
>> There are a few companies that have captured an unbelievable amount of value by releasing really radically new technologies. So open AI, anthropic um and then sort of connected to that Microsoft and then there are these chip companies like Nvidia that make enabling technologies that allow these techn these large language model tech based techs to to emerge. That's so that's the new world we live in and the market has been told over and over again that this is absolutely going to change everything about the world and they have continued to boost these um relatively small portfolio of stocks assuming that the miracle is going to happen and it's going to go up and up and we are going to have an unbelievable payday because of these new technologies.
>> Right? There's a lot to unpack there and it's not only Sam Alman who's talking about a uh a possible AI bubble. We're hearing this from Jeff Bezos as well who says that the current AI investment surge resembles a quote industrial bubble. We know that powering and developing AI is notoriously costly. Right? We've seen how much money is being poured into AI from new startups to these massive data centers. Um, and companies like Nvidia, which you just mentioned, which makes the chips that power most AI tools, are now being valued in the trillions of dollars, these astronomical sums. So, for those of us who don't know how to read the tea leaves here, Paul, um, what are the warning signs here? How do I how do we identify these warning signs? What should we be looking out for that's giving so many people in and out of the tech industry? This idea that the AI boom might be peaking.
>> Well, I think the tech industry just kind of gets up to this like it did with blockchain and just tells you that everything's going to change and if you miss it, you're a fool. The bet that the market has made is that all the value is going to get caught by like like five or six companies and they're just enormous like Nvidia, right? And so for that to happen, it means that they have created all the value and they're going to be able to hold all the value. And um in general, the way the tech works, especially internet technology, it's really hard to keep all the value for yourself. You know, Microsoft tried with the web in the early days, and people have tried to, you know, when something changes, when something is big and new, people try to get in there and really kind of own it and control it. But the the history of the internet and the history of the stuff is that if there is something there, and I think there is, there's definitely plenty there, it gets spread out. And so even if we we probably are in a bubble, which is my instinct because we're kind of always in a bubble in this industry. There's always a little more froth and a little more venture capital going than than anybody's ready for. But even if there is this real value in the in the technology, it's going to need to be distributed in terms of where people are investing, what their portfolios look like, and what they're doing as opposed to all of it locking into these few industries. If that does happen, then we're not looking at a lot of innovation in the future. We're not looking at a lot of growth. We're just looking at what uh Jensen Hang and and and Jeff Bezos and Samman think we should be looking at.
You know, it's interesting that you say, and you're not being hyperbolic when you say that, you know, a handful of companies hold all this value. We're talking about approximately 80% of US stock market gains in so far this year in 2025 are attributed to a handful of AI related companies. Among them, Nvidia, uh, Meta, Alphabet, the parent company of Google. And I'm sure by now you've seen this uh graphic that I believe was first published by Bloomberg which shows sort of Nvidia in the center, Nvidia pouring cash into other AI companies. Those companies then turn around and use that same cash to buy Nvidia chips. Isn't this type of sort of circular financing also a feature of a bubble?
Manny, you don't see that as like a beautiful everybody's holding hands, you know, you know, circular economy. They did it. It's No. Yeah, of course. Right. Like there there there's a lot of self-deing. Some of this is just always really normal with big business. Like, you know, it's there's all sorts of strange economics at work. When when giants get together and make love, it's terrifying to the rest of us. But but in this case, boy, it is suspicious. It's also because they're they're speculating, right? like, I'm going to get this many chips from you because if I get that many chips, I'm going to get to do magic tricks and the magic tricks are going to make so much money that I'm going to need more chips. And so, like, could some of that be real? Sure. Does it look like a lot of really powerful people sort of making sure that all the money stays in the bank vault in the basement? Yes. So, yeah. I mean, look, when you get that that weird feeling in your tummy, like, yeah, there's something up. like it's not
>> I I love a lot of these tools. I love technology. I love what they're doing. But yes, when when billions of dollars are getting moved without actually getting moved, um you got to start looking a little more closely, right?
No. What's also true is that the conventional wisdom of the United States right now is that AI is an arms race. If China can build competitive AI systems at a fraction of the cost, could that undercut the valuations of US tech giants and push the AI bubble closer to bursting?
>> We saw the the market really dip when DeepS announced its um a version of its LLM that came out and they trained it much more quickly and much more cheaply using less expensive GPU chips. Right? So because there's there's a lot of embargos as to what China can use out of Nvidia and similar companies and so the market tanked for a day and of course it it merrily went back up again but people they broke out because the promise had been that we had this amazing barrier and you know the other thing you're seeing is like it's often quite hard to differentiate between chat GPT and claude's output. Some of the open- source models might be hard to run or less available, but they're from China, but but when you use them, they're very good and very effective. And often they're cheaper to use and cheaper to access. And China, you know, I mean, the the deep sea folks are are roughly aligned with the mission of the Chinese government. And so there's a different kind of America's like, we'll put more fuel in the tank, whatever it takes, just build data centers. Every it's all going to work out because of capitalism. And China is more like, well, we're going to need to compete and we're going to need to do it this way and we're going to put we're going to be very strategic. But what they showed and I think what really blew everybody's mind is the narrative was one of like true great American genius. And there has been true great American genius in standing up this technology, but the moat got so much smaller so quickly and nobody was ready for that. We just had this story of like total triumphalism around this new technology and everybody else was kind of limping along including Europe and then China was just like actually how about cheap and fast maybe a little derivative of you but if you if you read um interviews with the founder of that company one of his goals is to compete is to say China can do this too he is patriotic and proud and he wants to show us what's up.
>> So Paul let's talk about those who are not buying the hype. I want to point to Michael Bur, the investor who famously predicted the 2008 housing crash, has made new bets that AI giants like Nvidia and Palunteer are overvalued. That's turning a lot of heads. What does this tell you?
>> You know, it's funny because when everybody got really overheated around crypto, it wasn't touching the real economy in every single way. And so the various crashes and ups and downs in crypto world even though the west coast and and sort of Silicon Valley was really really excited about it. It didn't seem like they came home like it didn't feel like it was going to blow up your mutual fund. This one we are all in. It is sort of embedded in everything and so it's scarier, right? I I got to tell you I'll tell you my instinct from there's so much value here. Someone uses this technology. It has been misused. It has been misold. It has been abused. It is very dangerous for society. But it can do things that used to take 40 people and it can, you know, 6 months and sometimes it can do them in 3 minutes. And I think about this as a technologist. So obviously there's something a little explosive here. It could be very dangerous for the economy. It could cost a lot of jobs, but there's there's something real, right? But um but at the same time, yeah, I mean, is it it's a pretty safe bet to say, whoa, Nvidia grew really really fast. I don't know if they can sustain that. Maybe OpenAI needs that much. You know, they at one point Microsoft bought three Mile Island because they needed they thought they would want nuclear power for their data center. So, so that push is still going. The data centers need to get built in order to do this stuff. It can be an ecological disaster. They need nuclear power. It's all too much and they're wasting a lot of time and energy. Nvidia is overvalued, but maybe 15 years from now, it won't have been. So, I think we're all triangulating around that. I will say personally, my co-founder and I are I'm I don't own a lot of uh shares. My stuff is in my business. I believe in a lot of this stuff, but like I would not go out and buy Nvidia stock because I just don't know. Like I just I cannot, you know, I'm not a financial adviser by any means at all, but I'm like I would pull back from that stuff because it really does feel overheated.
More with Paul after the break.
So Paul, if the AI bubble does burst, what's really at stake for the industry, for the world economy, and for those of us using this new technology every day?
Well, look, there's there's two kinds of burst. There's we were totally wrong and the technology didn't work and we lied. Okay, which I actually think is true of a lot of blockchain applications. They worked in theory, but in practice, I'm still buying things often with cash, right? Like it just like didn't take over the world in the way that people thought it would. Um, so there there's that. There's like it's it was all a lie all along. It doesn't really work. I don't think that's the case in this one. So, we're we're going to be left. Let's say the bubble does burst. Friend of mine once said during one of the many crashes I've lived through, he just turned to me. He said, "A lot of money is going to money heaven today." And that's that's what happens. And and this one's going to if it does happen, it's going to sting because it is all the way through the economy and it is all the way through the you know the index funds and it is all the way through the the the retirement savings of people. And I don't think it's like I don't know. It could set off it's could set off all sorts of downstream effects.
And >> well, we certainly don't want to sound alarmist, but this is the concern. The concern is that 2/3 roughly of Americans own stocks. Those accounts could be deeply affected, retirement accounts could be deeply affected by some estimates. We're talking about trillions of dollars held by American households that could be wiped up, as you put it, to money heaven. I do think it's not I mean it's not like you know 2008 was hopes that is the American economy and so that to to have real estate completely crash was that's like finding out like dirt isn't there anymore you can't get water like it was the it's as fundamental as can be. This is several layers up. The tech industry obviously works. We're going to continue to buy things through Amazon. So I wouldn't expect a meteor to hit the economy in quite the same way. I I'll be very frank. The way I I look at it is obviously things will have to balance out because you just can't have these like a couple giants holding on to everything. It doesn't work. Like stuff just slips through their fingers. Does that change in allocation? Does it mean that all the bets the VCs are making on startups now does some of those startups kind of break through and they become also sort of participants in the you know publicly shared economy and are they are they something that you buy stock in and it sort of balances out over time and okay yeah Nvidia had a bad quarter but there's these other things going on or is it sort of like the the internet crash where you just have this bad dip now eventually things come back roaring and
>> I mean you know what whatever you think of Amazon it's a a big company, right?
>> Listen, Paul, this is what I was going to ask you next. You know, we're talking about 2008 and and the housing crisis. But let's back up a little bit because the last time that we saw this kind of excitement over a new technology from the market like we're seeing right now with AI was during the dot boom in the '90s.
>> Millions of Americans own a personal computer. If you're one of them, you can now glimpse the future with nothing more than a modem, a phone line, and a few dollars a month.
that also led to market correction. Some people who uh follow tech say that these two moments do feel uh eerily similar. Can you break down why this bubble matters to ordinary people and what lessons um we could take from the dotcom bubble?
>> The market is a funny thing, right? Like it believes so much right now and people believe you you've got your fear and your greed and right now greed is winning but fear might show back up. When fear shows back up, you can look at it as a really sad moment in which money goes to money heaven and we have to go to the funeral home for money and stand around the coffin and say goodbye. Or you can say, "Wow, it's all a little bit more open now. It's a little cleared up." But what I remember from that area era before I I remember it was crooked. Like it just started to everything felt like a scam. Somebody told me that my blog would they could sell shares in it through a boiler room situation and I could get a credit card in the in the you know in the island and I was just and I you know I'm like I don't understand like I was just like don't talk to me about that right and so if you care about real value and you want to help people and you want to do things there's going to be lots of tech industry left over I hope so those are my memories my memories are not of this great tragedy for the tech industry but it is tough If people were leveraged and that they have made this bet, I would be very, you know, you you always have to hed your bets, which in America means you get a lot of fixed income and you get a lot of bonds and you you balance it out with this stuff. So, is the entire economy going to go? Probably not with this one. It just doesn't feel it's just a bigger economy and there's still a lot of fundamentals. We're a little too online for it to all go away again.
One thing that we hear from economists and and from experts, you know, not tech insiders, but economists looking at this from the outside is that oftentimes with bubbles, you don't know you're in a bubble until it's it's too late. And when we look back at moments like the dot bubble bursting 20201, that was very painful for for many people. About $5 trillion in the market was wiped out. The NASDAQ fell by 80%. It took 15 years to fully recover. So people who remember, I was a teenager back then, but people who remember are looking at this moment and I think trying to look at the signs. What looks familiar, what looks like a red flag, what looks like a warning. And I think for a lot of people right now, the what's threatening about AI is the idea of job loss, of layoffs, not in a single industry, but across many, many industries. And it's one of those things that an AI bubble could potentially set off. But the truth is layoffs have already started without the burst. And the irony is jobs seem to be at risk either way, right? Companies are cutting positions as they invest in automation. But if the AI bubble does burst, that could lead to more layoffs, too. So what does what do you think the hype around AI replacing jobs gets right? And what is it missing in the moment?
Well, first of all, I think a lot of the the layoffs that have been going on over the last few months have been AI has been kind of dropped into the narrative, but a lot of them are just people dealing with the overhiring that happened a few years ago and not really having a plan for all the people they brought in. They just assumed a specific kind of growth that didn't happen and big corporations do this. Um, it's awful, but there it is. At the same time, some of the recent advances have definitely thrown me for a loop. I got to be frank. Um, and very recently things like cloud code and and open AI's codeex are doing work that I used to do years ago and they're doing it really well and they're doing it quickly. I don't know if you necessarily see everybody getting sent home. The question is how does growth happen, right? like like things are going well for your company and then you would typically hire 10 new engineers because you need to build out your Salesforce installation and you know all those things are we we've sort of built that into the system and you just may not need that anymore. You may just be able to like get a smart person in the corner to kind of glue some stuff together with the Microsoft AI tools and and they're just going to be pretty happy with that. and you previously that would have been five people and now it's no people. So I don't think we've even started to have that conversation. I don't think that our current system of government is ready to have that conversation. I don't think our billionaire um company owners are particularly interested in that one except in a kind of general market pleasing way. And it definitely leaves me real squirly cuz I don't I used to really think like, hey, we're going to find a path, but I think there might be more shocks ahead. And and I wish I could say something different.
>> You know, the the pace at which the technology continues to evolve and improve is is shocking still to a lot of people. I mean, we should note that this month marks 3 years since Chat GPT's release to the public. We've seen new models of Chat GPT launching regularly. The rate of these new iterations of the LLM is growing exponentially faster. In that article that you wrote for Wired, you wrote that you hope that AI will eventually become normal. You said that you love normal technologies. I'm hoping that you can tell us what you mean by that and what would it take for AI to actually get there.
>> Manny, I am a tremendously boring person and I like when things are boring and stable and predictable so that you can build a business or or know what's going to happen next. You've got two things. First of all, you had 3 years of AI company founders telling you that it might emerge into a galactic star baby at any moment. And the press loved that and it was really overwhelming. And that was a lot for people to process and they're kind they've kind of backed off of that. Then you have these enormous step changes in what it's capable of. And so the original versions of chat GPT it'd be like you know count the RS and strawberry and it would do a terrible job. It's all getting a lot smarter. It's getting a lot better. And so what's actually happening in the industry and I think it is becoming more normal. The exponential changes in the capabilities of the models that that rate has slowed down. But what is increasingly accelerating is the sort of classic tech industry product that is growing up from these models. So there's more and more actually useful stuff and less and less experimental stuff that is happening. This weekend I was playing around cuz I like to keep my hands dirty and I like to learn and I built four apps and some of them would have taken me months to get done in the past and four is too many. Like that that's too much for a weekend. So, we're going to have to ride that wave because you kind of can't put it back in the box. And at the same time, it's starting to feel more familiar, just really a lot faster. It's no longer feels like it's they're no longer trying to convince us that it's a consciousness. They're trying to remind us that it's software. And software is familiar. We all know software. You know how to, you know, edit text and open Google Docs and and sort of do all those things. And I I for me the hope is always that it will fit back into that world so we can talk about it. And so for me when we can just finally have that conversation and we're not even ready for it yet. I know that. I accept that. But when we can have that conversation, I'm really excited. I'm ready. Uh cuz then we can figure out what we're going to build long term.
Paul Ford, thank you so much for joining us today on The Take.
>> What a joy. Thank you.
[Music]
[Music]
[Music]