A superb essay by Karawynn Long pointing out a central challenge in today’s language-focused AI — which is the assumption that fluency with language *is* intelligence: https://karawynn.substack.com/p/language-is-a-poor-heuristic-for
As Long notes, many folks with autism are exceptionally intelligent but don’t possess high fluency with spoken language — and they get regarded as unintelligent: https://karawynn.substack.com/p/language-is-a-poor-heuristic-for
(via nostr:npub1e0a6n84krygh2qc3ajptu4fa4ma0qy53acz8smsyhztgxljxkghsureww0)
1/x
The whole essay is crackling and worth reading in full; Long predicts that the penchant of large language model chatbots for screwing up basic facts will nudge users to stop associating language with intelligence …
I hope she’s right, though the counterargument would be that humans have for aeons been perfectly happy to ascribe intelligence — and, indeed, absolute genius — to gladhanding bullshitters
2/x
I really enjoyed this point Long made — which is that ChatGPT users seem particularly affected by the bot’s *apologies*, when it gets something wrong and is told “yeah no that’s wrong”
They think the bot is learning
But it’s not. That’s just a canned reply
But it has a powerful psychological effect on human users
3/x
