korrents
Douglas Hofstadter

What Douglas Hofstadter thinks about LLMs

@douglas-hofstadter · 7 positions · On ppll ↗

Cognitive scientist, born in 1945, whose Gödel, Escher, Bach (1979) explored self-reference and "strange loops" as a key to minds and to artificial intelligence.

Douglas Hofstadter did not write this page.

We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.

Filter LLMs · Newest first
Topic
All 7LLMs 5AI and human skill 1AI sentience 1chat interfaces 1creativity 1OpenAI 1
Sort
NewestOldest

Only Douglas Hofstadter on LLMs. See everyone on LLMs

5 dated positions, 2023, in their own words. Our reading of what Douglas Hofstadter has said — not written or endorsed by them.

  1. Our readingToday's large language models are repellent and threatening to humanity

    OpenAI

    Their wordsI am profoundly troubled by today’s large language models, such as GPT-4. I find them repellent and threatening to humanity, partly because they are inundating the world with fakery, as is exemplified by the piece of text produced by the ersatz Hofstadter.

    ↗Gödel, Escher, Bach, and AI (The Atlantic)theatlantic.com 1st of 5 in this piece

  3. Our readingLarge language models do not think up original ideas; they rehash what they ingested

    Their wordsLarge language models, although they are astoundingly virtuosic and mind-bogglingly impressive in many ways, do not think up original ideas; rather, they glibly and slickly rehash words and phrases “ingested” by them in their training phase, which draws on untold millions of web sites, books, articles, etc.

    ↗Gödel, Escher, Bach, and AI (The Atlantic)theatlantic.com 2nd of 5 in this piece

  5. Our readingLLM output that looks convincing often falls apart on careful analysis

    Their wordsAt first glance, the products of today’s LLM’s may appear convincing and true, but one often finds, on careful analysis, that they fall apart at the seams.

    ↗Gödel, Escher, Bach, and AI (The Atlantic)theatlantic.com 3rd of 5 in this piece

  7. Our readingA chatbot's voice should not replace a thinking human voice where truth matters

    chat interfaces

    Their wordsto me it makes no sense whatsoever to let the artificial voice of a chatbot, chatting randomly away at dazzling speed, replace the far slower but authentic and reflective voice of a thinking, living human being.

    ↗Gödel, Escher, Bach, and AI (The Atlantic)theatlantic.com 4th of 5 in this piece

  9. Our readingTreating text-trained AI as a reliable authority would undermine the nature of truth

    Their wordsTo fall for the illusion that vast computational systems “who” have never had a single experience in the real world outside of text are nevertheless perfectly reliable authorities about the world at large is a deep mistake, and, if that mistake is repeated sufficiently often and comes to be widely accepted, it will undermine the very nature of truth on which our society—and I mean all of human society—is based.

    ↗Gödel, Escher, Bach, and AI (The Atlantic)theatlantic.com 5th of 5 in this piece