Episode

"theory uplift differentially benefits safety & is massively underpriced" by Yudhister Kumar

Podcast
LessWrong (Curated & Popular)
Published
May 27, 2026
Duration seconds
153
Processing state
not_requested
Canonical source
https://www.buzzsprout.com/2037297/episodes/19246944-theory-uplift-differentially-benefits-safety-is-massively-underpriced-by-yudhister-kumar.mp3
Audio
https://www.buzzsprout.com/2037297/episodes/19246944-theory-uplift-differentially-benefits-safety-is-massively-underpriced-by-yudhister-kumar.mp3
JSON
/v1/public/podcasts/lesswrong-curated-popular-5643401/episodes/theory-uplift-differentially-benefits-safety-is-massively-underpriced-by-yudhister-kumar
Markdown
/podcast/lesswrong-curated-popular-5643401/theory-uplift-differentially-benefits-safety-is-massively-underpriced-by-yudhister-kumar.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/lesswrong-curated-popular-5643401/episodes/theory-uplift-differentially-benefits-safety-is-massively-underpriced-by-yudhister-kumar/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/lesswrong-curated-popular-5643401/theory-uplift-differentially-benefits-safety-is-massively-underpriced-by-yudhister-kumar.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

[1] We will likely have near-superhuman mathematics AI by Q1 2027. [1] [2] Qualitatively, AI mathematics capabilities are developing significantly faster than automated AI R&D capabilities. [2] [3] Thus, we will likely have a period of time where the rate of our ability to rigorously & usefully verify and understand model behavior and model outputs outpaces the rate of capability development itself. [4] Our ability to take advantage of this period is bottlenecked on the qu...