I appreciate how democratic this is, in comparison with other platforms: people feel themselves free to directly tell the founder “this is a bad idea”, and I’m willing to believe that he has not enough power (or desire) to force it through. Props to Jimmy Wales - not for this dumb take, but for the whole system.
And I do think it’s a bad idea. How many new users, once seeing a reply written by a LLM, will immediately think “I wasted my time with this edit. My reward was to get some whiner who couldn’t even be arsed to write the complain sending a bot to whine against me? Fuck this shit.” and never submit a second edit? Fuck, myself dropped another wiki for smaller stuff.
Some editors there also mentioned LLM tendency to make shit up / assume policies out of nowhere; this gets really bad, because no help is less worse than misleading help.
LLMs also have a tendency to force exactly three elements, even when it would be better to list more or less of them; Jimbo Wales’ ChatGPT example shows it. (Also, sod off with those emojis, they add absolutely info, only distraction.)
EDIT - also note he’s trying to solve a problem, he isn’t just trying to shove AI “because FUCHURE!” crap.
This is a refreshingly civil and rational conversation about it though, sounds like he’s not trying to jump the gun
A.I editor? He’s talking about A.I giving more context as to why a page is rejected. It’s not editing the page, just filling in a part of a template that a human volunteer might not have time to.
Honestly, I don’t see a problem with that. Rejecting an edit or page without a why is like getting an error message with most stack trace. It’s not very useful.
