[yyyk]

Hmm. It misses the ironic 'thanks to HN advice' point, and more importantly it mixes up suspensions (if I understand correctly): the main ads account was not suspended again *. First the secondary ads account was suspended and only then the main ads account.

* This proves ChatGPT has no world model. If it had one, it would have been aware that undoing a Google suspension is impossible, therefore no account can be suspended twice. /j

[cinntaile]

I got -3 for asking that question, I guess there are more people that do text summarizing for a living than I expected on HN.

It missing the ironic point isn't that big of a deal in a summary tbh, I wouldn't have included that either. Your other point stands though, that's an important detail to get right ä.

[yyyk]

> I got -3 for asking that question, I guess there are more people that do text summarizing for a living than I expected on HN.

I think it's because people are leery of threads being taken over by GPT produced comments. But so long as the requests/GPT comments/ are clearly marked, not too many, and with some justification for the request, I don't mind.

[wruza]

This is much better than OP though. It’s always confusing when people jump from one thing to another, while renaming things on the fly or implying details that are obvious only if you already know the picture. What chatgpt is really good at is structured narrative, even if it misses a couple of details here and there.