[jstummbillig]

There is hubris, but calling it a bubble simply does not check out, for one reason alone: If AI did absolutely nothing from here on out but give maybe a somewhat better version of current claude code (and it confuses me to no end that some people still refuse to see what is going on there, which admittedly are increasingly few of those who try, which makes sense because stuff gets better) that leads to, say, a ~2x dev speed up it, given the size of the market and how much software is missing still, AI as a whole would still be undervalued.

Of course, assuming that this would be the only thing where economic gains come from is already such a laughably bearish vision. It's just that that's all you need for the bubble-thesis to fall flat.

[marcyb5st]

This raises the question: hasn't the market already priced in a ubiquitous AI future? The current valuations seem to reflect the assumption that AI won't just 2x developer productivity but will also automate a huge portion of the workforce/boot productivity across the board. And I believe it has, and it's the only way to explain the sky-high valuations for companies that are 1) still losing money and, more importantly, 2) have no moat.

If that's true, then we are in a bubble by definition. When AI development eventually stagnates, failing to deliver on these promises, valuations will correct fast (and painfully). What happens then to Nvidia and other hardware companies? And what about the massive AI investments currently propping up the economy [1]? These would also be slashed, messing up the entire supply chain that's gearing up to meet this demand.

While I agree the technology is great and useful, I believe we are in bubble territory. I believe it's unlikely to be as transformative as the CEOs and VCs funding these companies claim.

[1] https://sherwood.news/markets/the-ai-spending-boom-is-eating...

[rurp]

A 100% increase in software development velocity is a wild claim to assert at this stage. The industry as a whole is not seeing anything like that and the handful of people making claims to that end are executives with a vested interest in boosting the hype. LLMs probably increase prototype or ad-hoc script development by that amount, but that only accounts for a minuscule amount of actual work being done in the industry.

Also whatever LLM productivity gains are currently happening are being massively subsidized. Once companies switch out of lighting money on fire mode most of these products will get dramatically worse and more expensive. Maintaining a cutting edge LLM isn't a railroad that you build once and can run and manage for centuries at a fraction of the initial price, they require constant expensive investment.