The Delphi Podcast

Jake Brukhman: Is ChatGPT Alive, OpenAI vs Stability AI’s Approaches to Avoid Killer Terminators, and Scanning Eyeballs with Worldcoin

Jun 6, 2023
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

LLM Capabilities

  • Large language models (LLMs) excel at statistically completing text based on existing data.
  • They are not necessarily human, sentient, or truly intelligent in the same way as humans.
INSIGHT

Emergent Intelligence

  • Predicting the next word in an LLM doesn’t just create statistically probable outputs.
  • It actually fosters a form of emergent intelligence within the model.
ANECDOTE

Detective Novel Example

  • Jake Brukhman uses the example of an LLM completing a detective novel.
  • To correctly identify the murderer, the model needs to understand the story's logic.
Get the Snipd Podcast app to discover more snips from this episode
Get the app