logoalt Hacker News

_aavaa_today at 1:56 PM2 repliesview on HN

Seems interesting, but buried under a mountain of written slop.


Replies

rightbytetoday at 3:20 PM

It is marketing for some SaaS thing. However, I think something like this, "verification loops with some sort of tricks", is what Anthropic and "OpenAI" do behind the scene with their interfaces.

sozaltoday at 2:00 PM

Fair on the length. The verification loop took DeepSeek from ~20 to ~80 on Web-Bench, matching Opus at ~7x less cost. Curious if the method holds up for you.

show 1 reply