logoalt Hacker News

pcuarontoday at 4:22 AM2 repliesview on HN

Excellent, so Gemini 3.7 can trash your codebase even faster (but seriously: don't trust this model)


Replies

drusepthtoday at 5:10 AM

What should you be looking out for as bad behavior from this model? I've daily-driven 3.5, 3.6, and now 3.7 for however long it's been out (alongside Opus, as I hit daily limits on both subs each day) and haven't noticed anything "untrustworthy".

seanmcdirmidtoday at 4:25 AM

Are 3.5, 3.6 any better? It is pretty flaky, Deepseek v4 flash is more reliable than 3.7 flash at least.