Fullstack developer shipping production mobile apps.

Mistral released open weights for its 1T-parameter Large 4. Alphabet fell about 12% over AI costs, Thrive bought ADT for $15.3B, and Reddit and Deepgram moved to control data and infrastructure.

Google limited its new frontier model Gemini 4 Carbon to cyber-defense partners, opened Search to MCP apps and expanded AI Mode ads globally; Anthropic and Nvidia shipped spend and speech tools.

Delaware's plan for AI-run companies drew a congressional probe, Bank of America told customers they own their agents' actions, and parents sued OpenAI over GPT-4o medical advice. Google launched Genie 4.

A harness cut 3,827 of 11,189 tool results into broken JSON while using 5.75% of the context window. If your agent seems oddly blind, audit truncation before blaming the model.

Across 147,002 coding-agent traces, early-failure predictors were at chance within a task (AUC 0.50), and plain continuation beat restart on solve rate and cost. Don't abort on prefix signals.

Across 3,000 turns, five models in Codex and Claude Code churned up to 31x the original code, and often kept editing after declaring the task done. Cap loops by diff budget, not the agent's verdict.