Stuff They Don't Want You To Know

Governments and AI, Chapter One: New Tools, Old Game

Sep 10, 2025
Governments are racing to harness AI, but at what cost? Explore the ethical dilemmas that arise as old governance practices meet cutting-edge technology. Revel in humorous critiques of AI's portrayal in popular culture versus reality. Discover surprising insights about hygiene in our homes while pondering the implications for government policy and public safety. This intriguing discussion balances excitement and skepticism about the future of AI in our world, leaving listeners both entertained and contemplative.
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
INSIGHT

LLMs Are Pattern Recognizers

  • Large language models (LLMs) are best described as advanced pattern-recognition systems rather than true intelligence.
  • Ben Bowlin and Noel Brown emphasize LLMs mirror human inputs and produce outputs from aggregated internet
INSIGHT

Fiction Distorts AI Expectations

  • The hosts argue fiction's sentient AI tropes mislead public expectations about real systems.
  • Real models are 'super vacuum cleaners' that assemble outputs from prior inputs and can hallucinate confidently.
INSIGHT

Better Filtering Drives AI Gains

  • Improvements in modern AI often come from better filtering and prioritizing of inputs rather than mystical leaps.
  • Noel Brown and Matt Frederick note sifting and discounting irrelevant signals drives exponential gains.
Get the Snipd Podcast app to discover more snips from this episode
Get the app