← @bgSponge

The Growing Gap Between ChatGPT and the AI Inside OpenAI - Noam Brown

Summary

Noam Brown tells Dwarkesh that frontier release cycles (often every two months or faster) are colliding with models that can already run week-long tasks and will soon stretch toward month- and three-month horizons.

  1. Release cadence outruns horizon eval.

    Two-month ship cycles cannot cover three-month agent horizons, so full-capability safety and product tests never finish before the next model.

    you don't have a way to evaluate the models at the full length of their capabilities

    tremendous concentration of power by the end of the year

  2. GPT-4-era policies are stale.

    Safety programs still assume short eval windows and have not been updated for long-horizon agents inside or outside the labs.

  3. RSI can freeze external releases.

    If three months of progress collapses into one, labs may skip classifiers and public deploys and recurse internally, concentrating capability.

  4. Math is the early gap.

    An unpublished internal model already solves hard and previously open problems the public cannot reach, creating an unfair external lag.

    math is the first domain where we're going to see that we're seeing this pretty clearly

  5. Both slowdown and secrecy have costs.

    Delaying releases helps eval; it also deepens the lab-vs-world disparity, and Brown says he has no settled weighing of those trade-offs.

    I don't have an answer for like how to weigh those trade-offs appropriately

Read the original ↗Save to your Sponge

Summarized with Sponge. Quotes are the author’s own words.

More from @bg

All readings →