m101

I agree with this sentiment entirely. Too much cheap talk. If you want to talk then you should be taking a position.

Perhaps these models just aren't as good as they say they are. If I had AGI I wouldn't sell the tokens. I think the public is helping them bootstrap to intelligence, but they'll promptly shut the door on our access as soon as they achieve it.

Isn’t this what jev training is supposed to improve? They don’t train on predicting the next token but predicting the distribution of next tokens, or something like that.

so if this was true then why do we need see GPU rental prices increase and openrouter token prices fall, for deepseek and other open models?

It’s a feeling that I got from reading his stuff over the years. Asking the AI just now comes up with a number of examples where he himself admitted to it. I also found him increasingly disingenuous.

I feel like Scott Alexander sacrificed the truth on the alter of status a long time ago

My interpretation of this is something like: if LLMs are to be mega useful token counts need to increase by many orders of magnitude -> broad adoption (and spending) would require token costs to drop by many orders of magnitude -> before the common folk get mega useful tools existing GPUs will be worthless

“However, the new investor presentation shows close to $30bn annualised revenues in July.”

2.5bn in revenue for all of July. That is a disaster.

Squeaky bum time

Eventually, because this guy felt unsafe multiple members of the rest of the camp put their hands up as feeling unsafe. I call it liberal mind virus because of the “safety” focus. A normal person would see small challenges and discomforts as part and parcel of daily life, and they certainly wouldn’t have safety concerns as grown up (mostly) men from a small girl. Everything was also taken out of context too - they all knew her and had camped together for 10+ years. It was basically a group-think public lynching based on some false sense of morality.

See my other comment

Yes very possibly. If you assume your competition will always exist then I think it makes sense to increase prices as they do even if you have the best business model.

OpenAI were initially competing with Anthropic in $ value of tokens, and then number of tokens, and then the 5h limit removal meant you could buy multiple plans for huge continuous usage. This gravy train was never sustainable from a margin perspective and must have cost them billions.

It’s now going to be interesting how anthropic react - 5h limit and retain 15-20x $ value, or go the same was as OpenAI.