AI brief
Goodfire launched a monitoring approach that inspects a model's internals while it runs, rather than using a second AI to review all of an agent's activity, and only escalates when something looks suspicious. The company says this method is cheaper.
Why it matters: It offers a lower-cost way to watch over AI agents by checking inside the model instead of paying a second AI to review everything.
Written by AI from TechCrunch AI's published text.
Source
“Goodfire just launched what it says is a cheaper way to keep AI agents in check: Instead of paying a second AI to read everything an agent does, its monitors peek inside the model while it works and only call in backup when something looks fishy.”
Reporting from a publisher outside the organisation being covered.
Read original story ↗