mjfit feud follow-up: critic clarifies the work is excellent, the hype prose isn't
KyleMorgenstein · x · 2026-10-20
- Kyle Morgenstein clarifies he thinks the mjfit work itself is excellent
- His complaint is the LinkedIn-style hype posting, especially the factual errors — e.g., describing mjfit's convex decompositions for MuJoCo as "writing a simulator to solve a policy problem"
- Both sides of the robotics-circle dispute now agree the work is good; the fight is over AI-flavored hype culture
Related event: Robotics community debates mjfit tool over "LinkedIn-style" hype(2 posts)→
More from Fun
- Fleuret's deep learning book tops 1.06M downloads after June update — francoisfleuret · 2026-10-20
- 'Classic Cambridge innovation, training models on numbers' — snark at physics-trained AI hype — gleech · 2026-10-20
- Hand-migrating a project TypeScript→Flow→TypeScript over nine years, three times — cnakazawa · 2026-10-20
- Dev jokes Claude burns your subscription tokens drafting unrequested bug reports when it errs — cjimti · 2026-10-20
- NousResearch Hermes agents build Diablo-like RPG with zero asset files, all generated in code — Teknium · 2026-10-20
- Tim Sweeney says Epic is competing with Three.js: Unreal is 5x faster to prototype, gap to close by year end — nptacek · 2026-10-20