From ABC7 Chicago · · 1m
Anthropic's Claude builds next AI model
Anthropic's AI model Claude now leads 26% of its own research to create a smarter successor
In brief
Anthropic's AI model Claude now leads 26% of its own research to create a smarter successor Claude autonomously contributes to 26% of Anthropic's R&D work while staying under human supervision Originally reported by ABC7 Chicago.
Claude leads R&D
Anthropic's Claude model now handles 26% of its research and development work while remaining under human supervision
Progress over time
Claude's role grew from zero percent in February to 26 percent in August, showing increasing autonomy within human oversight
Human oversight
Anthropic maintains 30,000 agents for research and engineering work with monitoring systems to detect misbehavior
Safety concerns
Anthropic CEO Dario Amodei and tech leaders call for slowing AI development to address safety risks
Public transparency
Anthropic urges AI labs to share development metrics publicly to track progress and ensure human control
Recursive self-improvement
Anthropic warns that models building themselves could challenge human understanding and control of AI systems
Here's the gist
Claude autonomously contributes to 26% of Anthropic's R&D work while staying under human supervision
