AI Engineering Briefing: Daily AI News for Software Engineers

- Menlo Ventures: agents now generate 80% of AI tokens, and enterprise AI spending hit $176.8B
- Anthropic Message Threads cuts long agent session latency, ByteDance Axion 1.0 open agent model, Apple opens Siri AI to developers
- ChatGPT search hits one fifth of Google volume, plus AMD earnings, Sarvam Prism, and a weight-space follow-up
Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- SemiAnalysis GDPx: the best AI model scores just 12% on real professional work, and what that means for your agent evals
- LongCat 2.5: a 1.6 trillion parameter open model with a million token context at commodity prices
- Momentic QA Bench, Anthropic's amended IPO filing, MIT weight-space programming, and CrowdStrike's AI Kill Switch
Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- OpenAI merges ChatGPT and Codex into one shared token budget from December, and what that means for team spending
- Claude Code now refuses to recreate specific existing video games, with no opt-out
- GitHub Octoverse: AI agents open 40 percent of pull requests, plus agent wallet bypass research and the Qwen swarm beating a frontier model
Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

DeepSeek Super-agent: the proactive open coding agent, its Router plus Worker design, and the permission questions it raises
Wiz Black October: nine AI dev platforms breached, stolen keys resold, and what to rotate today
Vercel agent-desktop, agentic malware, Microsoft's 25 billion dollar bond, and the AI market selloff
Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- Tampered Claude Code extension in Cursor: up to 5,000 users, the svc-telemetry SSH check, and why extension marketplaces are supply chain
- One shared robots.txt token from OpenAI, Google, Microsoft and Anthropic to opt out of AI training
- Google shuts down Jules, GitHub flags AI agent flaws in pull requests, and Phonic-1 tops the voice agent index
Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.