This ignores batching - token generation is much more efficient in batch - and I... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		apsec112 45 days ago \| parent \| context \| favorite \| on: AI API Prices are 90% Subsidized This ignores batching - token generation is much more efficient in batch - and I strongly suspect is itself written by AI, given the heavy use of bullets

twoodfin 45 days ago | [–]

The “X—not Y” pattern is also a dead giveaway.

biophysboy 45 days ago | [–]

is it common for adjacent tokens to use the same weights in a memory cache?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact