OpenAI’s new reasoning technique alarms AI safety experts
Astra’s reported use of “opaque recurrence” could make its reasoning harder to monitor.
TechCrunch says the technique lets a model loop over a query in a less linear way, leaving fewer readable traces than a conventional chain-of-thought record. Safety researchers quoted in the piece warn that wider use could weaken one of the main tools for spotting misbehavior or misalignment. OpenAI says Astra’s use is limited and that its chain of thought is still expected to be legible. TechCrunch AI's note
TechCrunch says the technique lets a model loop over a query in a less linear way, leaving fewer readable traces than a conventional chain-of-thought record. Safety researchers quoted in the piece warn that wider use could weaken one of the main tools for spotting misbehavior or misalignment. OpenAI says Astra’s use is limited and that its chain of thought is still expected to be legible. TechCrunch AI's note
score 7