3 ms·
Maybe hackernews would benefit from an LLM that refuses appeal to authority type of arguments
by stabbles 3y ago
Maybe hackernews would benefit from an LLM that refuses appeal to authority type of arguments
- vacuity 3y agoIntroducing HackerNews Bot 9000 TM In light of dang's (Dan G.) unfortunate passing, we recognize the loss of a core member of the HN community and the invaluable moderation services that are now absent. We honor dang's commitment to upholding civilized, productive discussions. Although dang certainly can't be replicated perfectly, we hope the HackerNews Bot 9000 TM serves as a reasonable substitute while we look for a better approach. this is satire this is satire this is satire this is satire
- bamboozled 3y agoAttack the arguments, not the people.
- vacuity 3y agoNormally I would agree with that statement. I'm not sure if you say "people" to mean "proclaimed experts" or "people citing 'experts'", so I'll discuss both. For "proclaimed experts", there are often arguments on very uncertain topics such as this AI thread. When there isn't actually much in the way of concrete evidence, the kind that science is based on, the reputation of an "expert" may be a reasonable filter to avoid low quality discussion. Would that people don't take so much stock in these "experts" because the discussions themselves are full of glorified speculation, though. For "people citing 'experts'", someone who is citing "experts" that are of dubious background may not be arguing in good faith. At that juncture, it may be best to abandon the argument. I suppose that I don't think much of the discussions of AI here are productive because of the lack of concrete evidence and the slinging of personal beliefs as fact. They mostly come out as noise to me.