LLMs

  1. Post 10Analysis

    Why be uncertain?

    TypeSafe's Jev returns decisions with calibrated probabilities attached. Keeping those probabilities (rather than collapsing them to YES/NO) is what buys you composition, intelligent randomness, and a signal for when to try harder.

    Avery Louis 5 min read
  2. Post 2In the news

    Hallucinations as a new(ish) threat model for academics

    arxiv clamps down

    Josh Bloom 3 min read
  3. Post 1CompanyValency bond

    Hello World!

    Introducing Valency Bond, empowering people and AI to reason and ideate over the real corpus of cutting-edge research.

    Josh Bloom 4 min read