2025-10-21 08:15 UTC
@corbet@social.kernel.org I wish the author had skipped AI: Just be yourself! The authenticity and trustworthiness implied in this post is super-low.
On HN, there's some interaction with the author, indicating authenticity. But I think ultimately, the problem is that everything depends on interpretation of a single log line and an accusation that the company bricked the device as a retaliation.
I enjoyed all the tips on spotting AI-generated language: https://news.ycombinator.com/item?id=45503560
Replies (1)
-
@gnyman@infosec.exchange 2025-10-22 08:11
@benjaoming@social.data.coop @corbet@social.kernel.org Yeah, I've noticed that a few times when I try to use AI to "help" No matter the style guidance, it will use its own idioms and styles which make it sound very generic and boring. Which is exactly what one should expect. Hopefully it's just a phase, here was another example where the author shared his original "draft" which he asked the AI to improve, and everyone in the HN thread said the draft was nicer to read :-) https://mastodon.social/@daemontus/115411068479908761