Google DeepMind paper trains an agent to rewrite its own harness

(runtimewire.com)

4 points | by runtimewire 13 hours ago ago