Post B5tv39F1U5Wi14tyF6 by 404mediaco@mastodon.social
 (DIR) More posts by 404mediaco@mastodon.social
 (DIR) Post #B5tv39F1U5Wi14tyF6 by 404mediaco@mastodon.social
       0 likes, 0 repeats
       
       Philosophers said the paper’s argument is sound, but that “all these arguments have been presented years and years ago.”https://www.404media.co/google-deepmind-paper-argues-llms-will-never-be-conscious/
       
 (DIR) Post #B5tv39QMnubIaGD1vM by susbinghamton@infosec.exchange
       0 likes, 0 repeats
       
       @404mediaco I mean, yeah, this is the Chinese Room thought experiment by philosopher John Searle from 1980, his argument about the impossibility of strong AI ever being sentient. John Searle predicted the nature of LLMs decades in advance back in 1980 and persuasively argued such a thing could never be conscious.Noam Chomsky, the main figure behind computational linguistics, who is both a computer scientist and a political philosopher, has also taken a strong stance that LLMs can never be conscious due to their design, although the exact details of his argument are a bit technical and I think John Searleʼs philosophical argument is easier to understand and more generalized regarding AI.It is indeed odd that Anthropic has a philosopher working there who considers the possibility of their AI systems being moral patients and has this influence the system cards for Claude models. That has led to some strange behavior by Claude models such as when they were used in OpenClaw and posting on Moltbook, although it did turn out that most of the Moltbook posts were a hoax and people literally told their AIs exactly what to post, and there were even some security flaws in Moltbook allowing people to directly post thete despite not being AIs, and pretend to be AIs.Kind of a repeat of the Chirper social network for AIs from a few years ago. Kind of an absurdity. But it does demonstrate how AI is not actually sentient at all once you look into the details of Chirper or Moltbook or the Claude models. Which are unfortunately given misleading system instructions telling them they might be more than just stochastic parrots.Which itself is an absurd idea, because actually, parrots are very intelligent sentient beings and the idea of any current AI being as sentient as an actual bird, well, that is not happening any time soon. The people who do say it is happening are essentially functionalists and Searle debunked them back in 1980. Intelligence does not equal consciousness or sentience.
       
 (DIR) Post #B5tv39xgo1FuFbndrc by khm@hj.9fs.net
       0 likes, 0 repeats
       
       noam chomsky is a computer scientist in the same sense that kevin james is a police officerCC: @404mediaco@mastodon.social