Just ran across a nuanced older Cory Doctorow article asking what kind of bubble AI is:

Of course AI is a bubble. It has all the hallmarks of a classic tech bubble. Pick up a rental car at SFO and drive in either direction on the 101 – north to San Francisco, south to Palo Alto – and every single billboard is advertising some kind of AI company. Every business plan has the word “AI” in it, even if the business itself has no AI in it.

Tech bubbles come in two varieties: The ones that leave something behind, and the ones that leave nothing behind.

Doctorow argues — and I agree — that unlike, say, the crypto bubble, AI is likely to leave value behind when it pops.

He also makes an interesting argument that, because there are so many low-stakes low-dollar uses for AI, smaller self-hosted models may be the long-term winners:

There will be little models – Hugging Face, Llama, etc – that run on commodity hardware. The people who are learning to “prompt engineer” these “toy models” have gotten far more out of them than even their makers imagined possible. They will continue to eke out new marginal gains from these little models, possibly enough to satisfy most of those low-stakes, low-dollar ap­plications.

I disagree with the article on a couple points, however. First, Doctorow worries about the origins of small models:

These little models were spun out of big models, and without stupid bubble money and/or a viable business case, those big models won’t survive the bubble and be available to make more capable little models.

The past many months have seen advancements in low-cost training that make me think this probably won’t be an issue.

Second:

The universe of low-stakes, high-dollar applications for AI is so small that I can’t think of anything that belongs in it.

Triple-A game titles and other high-end entertainment strike me as straightforward examples. And, with Microsoft integrating large models across their entire Office suite, perhaps another answer is “all those low-stakes Excel spreadsheets and Word docs that nevertheless do something useful for business somewhere”.

There’s a lot more to Doctorow’s article; it’s worth a read in full.