Coding Ai

Can Rewriting an AI Agent Bend the Intelligence Curve? - Zhengyao Jiang: Weco co-founder Zhengyao Jiang discusses an experiment where an AI coding agent spent eight days rewriting the harness, code, prompts, and tools of another agent while keeping the underlying language model fixed. The discussion evaluates the reported gains of the resulting AIDE 85 over two years of human engineering.

Technology • Machine Learning Street Talk

Nvidia SoL-Pi system cuts coding agent token usage by up to 49%: Nvidia's SoL-Pi system reduces coding agents' token usage by up to 49% with minimal impact on performance by optimizing the control layer between the model and its environment. A research agent developed the system by testing 152 approaches across over 3,000 runs, though gains were smaller on other benchmarks.

AI Software • The Decoder

Microsoft launches redesigned Copilot super app and rebrands Scout to Autopilot: Microsoft has officially launched its redesigned Copilot "super app," which integrates chat, coding, and agents into a single interface. As part of the release, the company rebranded its Scout AI personal assistant, originally introduced at Build, to Autopilot.

Microsoft • The Verge AI

artificial intelligence

The Biggest AI Coding Agent Upgrade Is Already on Your Machine?!

Google DeepMind chief plans early release of Gemini 4 model: Google DeepMind's new chief, Koray Kavukcuoglu, plans to release Gemini 4 much earlier than the end of the year. The model is currently in post-training and is running internally within the Antigravity coding tool. Kavukcuoglu shifted focus toward trustworthy agents, calling AGI "not the right conversation."

AI Models • The Decoder