MDN using an LLM to generate wrong answers
2023-06-30 22:48:35.297908+02 by
Dan Lyke
3 comments
MDN can now automatically lie to people seeking technical information #9208
SummaryMDN's new "ai explain" button on code blocks generates human-like text that may be correct by happenstance, or may contain convincing falsehoods. this is a strange decision for a technical reference.
If I wanted vaguely human sounding autogenerated content, I'd use StackExchange.
Mozilla: Introducing AI Help: Your Trusted Companion for Web Development. Yeah, so it's a paid promotion.
[ related topics:
Interactive Drama Invention and Design Artificial Intelligence
]
comments in ascending chronological order (reverse):
#Comment Re: MDN using an LLM to generate wrong answers made: 2023-07-02 13:31:56.293181+02 by:
brainopener
That's fun. API design can now be driven by what's least likely to cause LLM's trouble.
#Comment Re: MDN using an LLM to generate wrong answers made: 2023-07-03 18:02:35.244159+02 by:
Dan Lyke
Yeah, between that and the outsourced labor force that's training these things realizing that they can automate their own jobs, we're gonna see some serious inbreeding.
#Comment Re: MDN using an LLM to generate wrong answers made: 2023-07-03 18:52:38.082646+02 by:
markd
digital hemophilia