4 ms·
Why would you ever consider asking an LLM a question like this though? They just make plausible-sounding things up, even if sometimes they're correct.
by spankalee 2y ago
Why would you ever consider asking an LLM a question like this though? They just make plausible-sounding things up, even if sometimes they're correct.
- bongodongobob 2y agoIf you tell it to use wolframalpha it could probably do some ballpark spherical cow physics.
- perrygeo 2y agoAmazing comment for two reasons: The idea of asking an LLM to generate code for a more rigorous but still high-level language is under-explored. This way you can experiment with AI - but at least have an intermediate code artifact that you can review and test before it gets published! Exposing the raw token stream is reckless. I had never heard of "spherical cow" science before but I love the imagery.
- Cyphase 2y agoYou really are one of today's lucky 10,000! For anyone else: https://en.wikipedia.org/wiki/Spherical_cow https://en.wikipedia.org/wiki/Spherical_cow Meta reference: https://xkcd.com/1053/ https://xkcd.com/1053/
- burnished 2y agoBecause with an LLM you can walk away confused for only a fraction of the frustration and time of the old way.
- ben_w 2y agoWhy? Same reason I'd ask a student in the second year of a degree — knows more than I do. > They just make plausible-sounding things up, even if sometimes they're correct. This is indistinguishable from my experience described above.
- mort96 2y agoI think the point of the comparison with LLMs was to disparage StackExchange, I don't think they disagree with your view of LLMs