>if model intelligence stalls. If model intelligence continues to improve

So much in AI is dependent on which of these two outcomes occur.

I am not so sure that's true. Even if AI intelligence were to plateau at today's levels, there are still many gains to be had in speeding up today's intelligences. ASICs with burned in weights could become economical to invest in as they would retain usefulness longer than 18 months, and I feel we have only scratched the surface on possible usecases of local AI and what it means for almost any technology product or interface.

Nice! I'm waiting for the day when the remainder of the Maya codices are solved by AI.

I was just thinking about the good old days combining human judgements on mturk for classifying high-value forum threads. Since replaced by using judgements from multiple LLMs.

The only use I can think of it these days is for social scientists who need to gather judgements from bona fide humans.

This will be devastating for a lot of workers in low-infrastructure countries.

And that it will be just another utility, like electricity.

This is like an earlier internet era when it was common for people to just copy-paste directly from other websites into a forum response.

Nice! But needs more footgun.

Needs more "invariant," its latest catchphrase

They're sweating it big time, and their fear is bleeding through their AI beneficence-speak. Time for the big US AI companies to call in some favors from the administration. But you can't unring a bell.

Recipes like this get baked into the next iteration of a model.

Well, that's silly. The batter goes in the pan; the recipe is supposed to stay out on the counter.

This is true, but when we compare them to OpenAI's repeated feints at altruism and nonprofit structure, the irony is plentiful.