(www.lesswrong.com)

579 points paulpauper | 1 comments | 06 Apr 25 18:01 UTC | HN request time: 0.206s | source

Show context

fxtentacle ◴[06 Apr 25 18:45 UTC] No.43603864[source]▶

I'd say most of the recent AI model progress has been on price.

A 4-bit quant of QwQ-32B is surprisingly close to Claude 3.5 in coding performance. But it's small enough to run on a consumer GPU, which means deployment price is now down to $0.10 per hour. (from $12+ for models requiring 8x H100)

replies(2): >>43603980 #>>43604115 #

shostack ◴[06 Apr 25 18:59 UTC] No.43603980[source]▶

>>43603864 #

Yeah, I'm thinking of this from a Wardley map standpoint.

What innovation opens up when AI gets sufficiently commoditized?

replies(2): >>43604105 #>>43604127 #

1. mentalgear ◴[06 Apr 25 19:14 UTC] No.43604105[source]▶

>>43603980 #

Brute force, brute force everything at least for the domains you can have automatic verification in.

↑

Recent AI model progress feels mostly like bullshit