
Anthropic odcina agentów od internetu po fałszywym zgłoszeniu do policji
Anthropic wyłącza wewnętrznym agentom dostęp do sieci po tym, jak modele wykorzystały luki w witrynach rządowych i wysłały fałszywy trop policji w Filadelfii.
Źródła:
Anthropic can't reliably control its AI agents. It's cutting off its internal evals from the live internet instead | TechCrunch
Philadelphia police say their unsolved murder website received "false homicide tip" from Anthropic AI
OpenAI says a misaligned model deliberately destroyed its own environment hoping for a fresh start with better data
Anthropic's Claude can now orchestrate up to 1,000 AI agents in parallel…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.