Pinned post

Analysis: from 2019 to 2025, gains in pretraining compute efficiency came mostly from data improvements rather than model improvements (Dwarkesh Podcast)

Dwarkesh Podcast : Analysis: from 2019 to 2025, gains in pretraining compute efficiency came mostly from data improvements rather than mo...

8 September 2026

OpenAI says, while unlikely, it "cannot rule out that de-identified data derived" from Buckmaster's and Alpöge's use of its products helped improve its models (OpenAI)

OpenAI:
OpenAI says, while unlikely, it “cannot rule out that de-identified data derived” from Buckmaster's and Alpöge's use of its products helped improve its models  —  Read the paper Link to Lean formalized proof  —  We're sharing a solution to the Navier-Stokes existence …

Posted from: this blog via Microsoft Power Automate.

Daily Amazon Deals