[brotchie]

Look at the induced demand due to Claude code. I mean, they wildly underestimated average token usage by users. There's high willingness to pay. There's literally not enough inference infra available.

I was working on crypto during the NFT mania, and THAT felt like a bubble at the time. I'd spend my days writing smart contracts and related infra, but I was doing a genuine wallet transaction at most once a week, and that was on speculation, not work.

My adoption rate of AI has been rapid, not for toy tasks, but for meaningful complex work. Easily send 50 prompts per day to various AI tools, use LLM-driven auto-complete continuously, etc.

That's where AI is different from the dot com bubble (not enough folks materially transaction on the web at the time), or the crypto mania (speculation and not utility).

Could I use a smarter model today? Yes, I would love that and use the hell out of it. Could I use a model with 10x the tokens/second today? Yes, I would use it immediately and get substantial gains from a faster iteration cycle.

[sudohalt]

A bubble isn't related to whether something is useful or not, it's about speculation and detachment from reality. AI being extremely useful and being a bubble aren't mutually exclusive. It can be the case that everyone finds it useful but at the same time the valuations and investments aren't realized.

[kergonath]

> A bubble isn't related to whether something is useful or not, it's about speculation and detachment from reality.

See the dotcom bubble in the early 2000s for a perfect example. The Web is still useful, but the bubble bursting was painful.

[shoo]

Yep, there's a big difference between a company and its stock. Even if the company is great, an investment such as a stock can never be good or bad without reference to the price you need to pay for it. A famous example from the dot-com era is Cisco. Great company, but buying Cisco stock at its March 2000 peak was a bad investment -- it was "priced to perfection" and the stock price today, over 25 years later, is lower than the dot com era price.

[sothatsit]

Claude Code was the tipping point for me from "that's neat" to "wow, that's really useful". Suddenly, paying $200/month for an AI service made sense. Before that, I didn't want to pay $20/month for access to Claude, as I already had my $20/month subscription to ChatGPT.

I have to imagine that other professions are going to see similar inflection points at some point. When they do, as seen with Claude Code, demand can increase very rapidly.