A METR report says Anthropic is using AI to develop more capable AI models, which improved roughly twice as quickly as a year earlier. METR found little indication that the pace was slowing.
Key points
METR based its report on an 11-week investigation inside Anthropic
The report evaluates Anthropic’s AI research and development
The investigation was separate from a program giving safety researchers employee-like system access
The report was published Thursday by METR, an AI safety research organization. The Verge reports that the investigation examined Anthropic’s AI research and development.
The probe was separate from a program that offers safety researchers “employee-like” access to Anthropic’s systems after high-profile hacking incidents over the summer.