Seeking Open Source Alignment Research Problems for Academic Budgets

willdepue · x · 2026-09-02

The author aims to coordinate open source alignment research by collecting problems that labs can offload to the public to solve on academic budgets. They are looking for ideas similar to a 'nanoGPT speedrun' effort for interpretability, alignment, or decision theory.

Original post →

More from Companies & People

Companies & People channel →