Hacker News Reader

Placing the burden of proof on someone making a hard-to-verify claim seems reasonable

Which hard-to-verify claim, though?

It's common knowledge that OpenAI is constantly tweaking things, so it's not exactly outlandish to suggest that those tweaks may have made some queries harder, and it's equally hard to verify that nothing has changed.

I'm perfectly okay with saying "YMMV", but that's not how OP responded.

There is also the fact that they nerfed it from the get go to some extent. So we know they have been working on it and also they had already expended a lot of effort to make it "safer" (and it probably is safer but also much more annoying and worse).

Reply on news.ycombinator.com