3 ms·

I guess we have to enjoy the last year or so before better models are indistinguishable from us? With some better/clever prompting I’m not sure they could be a

• smusamashah 2y ago
  I wouldnt mind a bot writing a comment on HN if it's thoughtful, which is HNs trademark. These comments are so damn void of any thought. You can flag them for being thoughtless alone.
  • dredmorbius 2y ago
    That's the premise of xkcd 810, "Constructive": <https://xkcd.com/810/ https://xkcd.com/810/> It's one of the more spectacular cases in which I think Munroe utterly missed the plot. It's not that it's possible to create bots which at a superficial assessment are making constructive / helpful comments. It's that bots which do this often have covert goals or purposes, "alignments" in the LLM/ML context. These might or might not be known or evident to either the bots' creators or users, and might even be exploited by still other parties. But bots are capable of a scale of operations which fundamentally changes the community dynamics of sites such as HN (or of Reddit, Twitter, Facebook, TikTok, etc., etc.). It also fundamentally violates the prime directive of HN, "[T]o make thoughtful comments. Thoughtful in both senses: civil and substantial": <https://news.ycombinator.com/newswelcome.html https://news.ycombinator.com/newswelcome.html> Or as dang so often notes, optimising for intellectual curiosity: <https://hn.algolia.com/?dateRange=all&page=0&prefix=true&query=by%3Adang%20intellectual%20curiousity&sort=byDate&type=comment https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...> Bots, so far as we're aware, are neither intellectual nor curious.