Elektrine lite

← Feed

@mxchara@cyberpunk.lol

2026-08-27 01:42 UTC

But if they're so confident that #Claude and #ChatGPT have surpassed the human brain then where are the head-to-head comparisons? These same geeklords talk about #IQ as if it were an infallible and empirically proven method of estimating not merely human intelligence but...everything, basically! They say IQ has immense predictive powers. (etc.)

Replies (1)

  • @mxchara@cyberpunk.lol 2026-08-27 01:44

    Hence it should be possible to assign an IQ score to #GenAI. Why hasn't that been done? Why are the "Generative AI" boosters so shy about running comparisons of their #LLM installations with human beings? Instead they merely run endless "benchmarks" comparing one corporate product to another. That's sufficient to convince the #tech geeks that their machines are constantly getting smarter. (&c.)

    Open ##4550270