Yacine mocks $2M RL env contracts that teach LLMs 'how to kill people'

yacineMTB · x · 2026-11-10

Yacine sarcastically calls out someone paid $2 million to build an RL environment teaching LLMs "how to kill people," who then acts surprised at how good the model is at it. The jab highlights the lurid task design in the booming RL-environment outsourcing economy.

Original post →

More from Fun

Fun channel →