Agents

tldr > evals are the new training data. instead of updating weights, you're updating the agent harness > problem is agents are famous cheate…

tldr > evals are the new training data. instead of updating weights, you're updating the agent harness > problem is agents are famous cheaters. they will reward-hack your evals and overfit just to mak

DGX agentx-post
agentsharrison-chase--x

tldr > evals are the new training data. instead of updating weights, you're updating the agent harness > problem is agents are famous cheaters. they will reward-hack your evals and overfit just to make the score go up > solution is treat evals like real ml. you need strict train/test splits and holdout sets to force actual generalization > quality over quantity. a few hand-curated, highly-tagged evals easily beat thousands of noisy ones > the better-harness loop is mine prod traces for failures -> turn into evals -> auto-tweak prompts & tools -> validate > the ultimate flywheel is more prod usage -> more failure traces -> more evals -> continuously self-improving agents

Related

Source: Harrison Chase (X) | 2026-04-10

Loading related sources…