There's this UK analyst claiming the AI financial boom right now. It's 17 times worse than the dot -com bubble. 17 times that number. Well, it really makes you stop and think, doesn't it? Deserves a proper look. It really does. Yeah. And thanks for sharing those sources with us. It's clear we're dealing with this massive contradiction, right? You've got these sky high AI valuations and then you look for the actual profitable uses right now. And it's well, it's a different picture.
Exactly. So today, that's what we're going to do. We'll unpack that bubble claim, see where AI is actually delivering ROI and also touch on this really intense work culture popping up again in Silicon Valley to try and meet. these huge expectations. OK, let's dig into that big number first. The 17 times bubble. This analyst, Julian Guerin, he's not mincing words. 17 times the dotcom era. Four times the 2008 crash. Beep.
That's a heavy statement. Mm hmm. And. It's crucial to get why he's saying that the dotcom thing that was built on maybe shaky revenue projections. Right. But this A .I. boom, he argues it's built more on just sheer market cap momentum. It's exploded way, way past any kind of near term profit reality. Yeah. You've got what, 10. major AI startups, together they've added something like a trillion dollars in market value. A trillion. But you look closer and almost all of them are
still, you know, deeply unprofitable. It's this profitability paradox. It really is. So you got to ask, who is making money here? And right now the answer seems pretty clear. It's mostly NVIDIA selling the hardware, the shovels for the gold rush. Basically, they're getting consistent returns. Pretty much everyone else developing the core AI models. They're basically just, well, bleeding cash, funding operations based on what they hope will happen down the line. And all that cash
burn. Yeah. It's making the traditional money people, the VCs, a bit nervous. You're seeing them pull back now. They look at the valuations and just think they're, well, kind of absurd. Right. So the whole ecosystem is increasingly propped up by just a few huge players. The mega backers, as Garen calls them. Yeah, not your usual VCs. We're talking SoftBank, sovereign wealth funds. Even NVIDIA itself is investing heavily back into the ecosystem. They're the
capital life support right now. And the bet seems to be if they just keep pouring money in, eventually one of these companies will crack artificial general intelligence or AGI. So Garen lays out two possible futures, basically. Option one. Yeah. The money keeps flowing, but it never really delivers that massive value everyone expects. It's just this capital treadmill or option two. Someone actually does achieve true AGI, a genuine breakthrough. And he's betting on option one,
isn't he? He seems to think the market hype has just gotten way out ahead of the actual tech capabilities right now. I'm going to say just trying to understand the tech itself can be tough. Honestly, sometimes I wrestle with prompt drift myself, you know, getting consistent results. So when you talk about these billion dollar market shifts. Yeah. Yeah. It's a lot to wrap your head around. So given all that cash burn right across the board. What's the single biggest factor keeping
these valuations so incredibly high? It really seems to boil down to those mega backers. Them plus this persistent belief, this hope that true AGI is just around the corner. OK, let's shift gears a bit away from the financials and towards some more practical stuff. If you're trying to get a better handle on the tech itself, Stanford actually has this really good four part lecture series out. It's about five and a half hours
total. Gives you a pretty solid kind of medium depth understanding of LLMs, you know, large language models, the AIs that generate human like text. And we are seeing some pretty cool jumps in what these things can do. There's a new tool, DeepSeq OCR, for reading documents. Apparently it's really impressing people. It managed to accurately read some properly messy handwritten. written letters, which is not easy. Yeah, that's huge. Imagine for archives or just
extracting data. But there's always this ethical tension simmering underneath, isn't there? Like Microsoft's AI chief coming out and publicly criticizing ChatGPT's ability to generate erotica. That feels kind of awkward given Microsoft put, what, $13 billion into open AI? It does. And you also mentioned the Australian government. suing Microsoft over confusion about AI pricing changes. It just shows how regulators are scrambling to keep up, trying to figure out how to handle
the ambiguity of these new AI systems. And here's something that feels genuinely concerning in the sources you shared. This trend of extreme overwork making a comeback in Silicon Valley. AI startups are apparently bringing back China's infamous 996 grind. That's 9 a .m. to 9 p .m. six days a week. Wow. Yeah, and there's even that dark joke going around among researchers about the 002 schedule, zero sleep, zero days
off, just two hours maybe. The pressure to deliver, to hit those short -term milestones, maybe justify those crazy valuations we talked about, it's pushing people really, really hard. It seems like a dangerous game. That kind of pressure cooker environment might get quick results, but... Burning out your best people. Yeah, and it introduces a huge risk of mistakes, right? Errors creeping into the very foundations of the models they're
building. Not good long term. Meanwhile, the big players are still pushing into new areas. OpenAI has apparently got a secret project, building a generative music tool. Ah, okay. Like SunoAI and those existing tools. Turning text prompts or even audio snippets into full songs. Always looking for that next big creative market to disrupt. So if this 996 work culture, this intense grind, is becoming widespread again, how does that really impact AI innovation in the long
run? Well, you might get some rapid gains, sure, short bursts of progress. But the risk of burnout, costly errors, and ultimately maybe less thoughtful, lower quality innovation seems pretty high. Okay, let's focus now on some specific tools and maybe more importantly, some really impactful applications. We saw Google talking about vibe coding. in their AI studio. Sounds like they're trying to streamline how engineers work, make development faster,
more intuitive. But the applications that really jumped out, I think, were in biodefense and health. There's this company, Valfos, backed by OpenAI folks. They launched with $30 million specifically to build AI biodefense systems. Yes, systems designed to detect and figure out how to counter new pathogens. In just hours. Yeah. Hours. I mean, think about that. That's a massive leap in how quickly we could respond to, say, a new
pandemic threat. It really is. And then there was that other study from Utah showing AI could spot parasites in human stool samples really quickly and accurately. Pause. Whoa. I mean, just imagine scaling that kind of detection speed, applying it globally to public health surveillance. That's tangible value. That justifies serious investment, you know? Absolutely. And the support
system for all this is growing, too. Amazon, OpenAI, NVIDIA, they're teaming up to turn Cal State into a big AI training center, recognizing we need more people who actually know how to build and use this stuff. There's even a paper mapping out like 90 different AI coding tools available right now. So thinking about all those tools and applications. Beyond the big names, what was the most surprising practical use case we came across this time? For me, it was that
health stuff. The AI detecting pathogens and parasites so quickly and accurately that speed just feels like a potential game changer for public health. Mineral sponsor read, insert sponsor read here. All right. So we've talked about the bubble fears, the potential, the pressures. Let's look at some hard data now, specifically from
industries already using this tech. We're pulling this from the Artificial Analysis 2025 State of Generative Media report, based on input from about 300 developers and creators using these tools right now. The data really highlights Google's position, especially in creative AI, for making images. Google Gemini is number one, 74 % usage among those surveyed. OpenAI's model, GPT image, is next at 64%. So, strong competition there. Yeah, and that lead carries over into video generation
too. Google VO. Leading the pack with 69 % usage in this group, competition's clearly driving things forward. And what's driving the choice for these creators? What makes them pick one model over another? Overwhelmingly, it's quality. That's the number one factor, cited by like 76 % to 82%, depending on the task. Makes sense. Cost comes second, yeah. But the message is clear. If the output isn't good enough, isn't high quality, then people just aren't going to use it regardless
of price. OK, so let's connect this back to that bubble debate we started with. This feels really important. Sixty five percent, almost two thirds of the organizations surveyed said they're already seeing a return on their investment in generative AI or they expect to see it within the next year. Right now. OK, that's likely self -reported data. You got to take it with maybe a small grain of salt. But still. 65%. That's pretty strong validation,
isn't it? It suggests that investment in these specific high quality focused applications like generative media is paying off for many right now. Yeah, it kind of counterbalances the hype around the purely research focused arms that are still burning cash. There's real measurable value being created in certain sectors. If quality is the absolute top priority, does that automatically mean the big expensive proprietary models from, say, Google and OpenAI will just always dominate
the open source alternatives? Well, quality is definitely driving adoption now. But the report suggests that the ability to fine tune models for specific needs and dealing with IP and licensing stuff, those are becoming really important secondary factors pretty quickly. So bringing it all together. What does this mean for you listening? We've really grappled with the central tension today, haven't we? On one hand, you've got this immense financial speculation that potential 17X bubble
Garen warned about. Huge risk. But on the other hand, you have clear, demonstrable ROI already happening in fields like generative media and these incredible potential breakthroughs in areas like biodefense. Yeah, it's a paradox, like we said. Massive financial risk running alongside absolutely essential functional progress. The debate isn't really if AI is valuable anymore, is it? It seems more about where the sustainable
value is actually landing. Is it with the companies selling the infrastructure like NVIDIA and Google through its cloud? Or is it eventually going to come from those research arms chasing AGI even though they're burning cash now? And maybe here's a final thought to chew on. If 55 % of organizations say they're already getting ROI from AI, What do the other 35 % need to figure out? What needs to change in their strategy or maybe the specific way they're applying AI so
they don't get left behind? That's the challenge of applying this knowledge in real time, isn't it? Absolutely. Well, thanks again for sharing the sources and letting us dive deep into all this with you today. Definitely check out those Stanford lectures if you're interested and maybe take a look at that generative media report data. Lots to think about. Out to your own music.