undefined | Better HN

0 pointsayewo4mo ago0 comments

Here's a recent comment [1] by an OpenAI engineer confirming that they do in fact make such trade offs between intelligence and efficiency.

[1]: https://news.ycombinator.com/item?id=46909905

0 comments

1 comments · 1 top-level

int_19h4mo ago

That comment only says that they have a lot of different options for smaller & faster models that people can opt into. It doesn't say that they dynamically scale things up or down depending on demand.

j / k navigate · click thread line to collapse