youtube.com web signal

Merchant: AI leaders can't stop fighting 'stochastic parrots'

TL;DR

  • Brian Merchant's October 2 Blood in the Machine post interviews Emily M. Bender about why 'stochastic parrots' angers AI industry figures.
  • The term comes from an influential 2021 paper by two AI researchers and two computational linguists, arguing LLMs just restate training data.
  • Merchant frames the backlash as financial: if LLMs are predictable, fallible systems, the AI boom's 'machine god' narrative loses its trillions.

"AI industry folks absolutely lose their minds over the term, and often bring it up, unprompted, just to try to debunk it again and again," writes Brian Merchant in a new Blood in the Machine post and companion YouTube video. The two words: "stochastic parrots."

The phrase comes from an influential 2021 paper written by two AI researchers and two computational linguists. One of them, Emily M. Bender, is Merchant's interview subject. The paper's claim is deflationary: large language models are "not 'intelligent' but are rather systems that output material based on what's in their training data."

Merchant's theory for the industry's reaction is bluntly financial. He writes that the term undercuts the very foundation that the AI boom is built on: "That tech companies are building a machine god of infinite power and potential." And then the follow-through. "If it turns out it's an unthinking system that's often fallible in predictable ways, well, that might not be worth quite as many trillions of dollars."

Shared on Bluesky by 1 AI expert

  • Emily M. Bender @emilymbender.bsky.social amplified

    @tommullaney.bsky.social

    "...We couldn't imagine that people would want lots and lots of synthetic text because it's just so patently useless." - @emilymbender.bsky.social on LLMs. #EduSky www.youtube.com/watch?v=7Z7o...

    View on Bluesky →