GPT-6 Astra experiment links AI agent learning failures to token budgets
Original titleFor example, GPT-6 Astra ran an experiment exploring why AI agents fail to learn with practice. It set AI agents’ token budgets too low. ...
AISummary
An experiment by GPT-6 Astra examined why AI agents fail to learn with practice, finding that setting agents' token budgets too low prevented learning. The researchers reported this sensitivity to the acquisition budget as a key finding rather than an error.
Source: Epoch AI · x.comPublished · added here