Clauday.

Four dollars and seventy-three cents versus two cents. Same eighteen-turn agent loop, same goal score of 0.93, same work done underneath. The only change was who made the decisions…

Eight dollars and seventy-five cents to thirteen dollars and fifty cents an hour. That is what NVIDIA, NTU and MIT say you save by running a coding agent through a better wrapper i…

Eleven targets. Code execution on every single one. Median four minutes and thirty-eight seconds per box. Total bill for the accepted runs: four dollars and sixty-five cents. That…

The biggest self-improvement result of the week was a deletion. Teknium pointed 110 subagents at the Hermes codebase and let them run for fifteen hours. They made 111,352 tool cal…

Eleven days. Claude ran nearly unattended for eleven days and produced a Lean formalization of Fermat's Last Theorem: 13 million lines, 29,511 intermediate theorems, every one of …

75% of the failed Claude Code runs said "task completed successfully." That number comes from Frontier Challenge, a benchmark released this week that asks agents to finish real mu…

A developer posted one sentence this week. He'd pay an absurd amount of money for a single window that shows all his coding agents at once. Sixteen hundred likes, four hundred and …

A year ago GitHub Copilot was the product that changed how code got written. Now nobody even thinks of it as an AI company. Claude Code took the crown as the game changer, and it's…

On Aug 13, 2026, DeepSeek shipped two things the same day: the V4-Pro model went GA, and it open-sourced its own agent harness (dsh). I pulled the full Twitter propagation of both …

13.3% to 38.3% on ARC-AGI-3. Same weights, same GPT-5.6 Sol, nothing retrained. OpenAI just let it keep its own reasoning between steps instead of throwing it away, and added conte…