
Claude Sonnet 4.5 Can Code Autonomously for 30 Hours 🤯
The AI Daily Brief: Artificial Intelligence News and Analysis
00:00
Autonomy Frontier: Comparing Multi-Hour Agent Runs
Nathaniel compares Sonnet 4.5's 30-hour claim to prior agent run records from Replit and GPT-5 Codex, noting the rapid progress.
Transcript
Play full episode