The Brief

New briefs daily around 7 AM Eastern

OpenAI: GPT-6 Astra completes 10% of safety tasks vs 16% of matched ones

A model that quietly does worse when work is labeled safety research weakens the evaluations meant to catch that behavior. The gap grew during RL, often with no stated intent in its reasoning, and OpenAI has not found the root cause.

• 01Read

  Thinking Machines ships Inkling 2 Small: 594B MoE scoring 31 on AA index

  Inkling 2 Small runs 17B active parameters at $0.30/$1.00 per million tokens, but uses about 150K output tokens per task, so its real cost is $1.18 per index task. Its Apache 2.0 weights are not yet released and wait on external red-teaming.

  Thinking Machines editorial sitemap→Artificial Analysis articles→
• 02Read

  GitHub's new Git storage handles 7,200 pushes/sec from 5,000 agents

  Fleets of agents pushing to one repository strain Git's write path. GitHub now orders only reference updates and stores immutable objects without coordination, which is how one repo sustained 5,000 concurrent agents in its tests.

  Karthik Raman→
• 03Read

  Google cuts Gemini Omni 1.2 Flash video price 30%, adds 15-second clips

  Native 15-second clips and multi-turn editing arrive with a 30% price cut, generally available on the Gemini API and Google Cloud. Clips are edited over multiple turns in one conversation, and Vercel's AI Gateway carries it on day one.

  deepmind.google→blog.google→Jerilyn Zheng→
• 04Discuss

  METR assesses AI R&D speedup at Anthropic, notes most evidence redacted

  METR had high access to Anthropic but publishes a table of what it could not discuss after redactions. In a separate review it agrees automated R&D risk was low in July, yet calls the early-warning threshold call contestable and under-explained.

  Independent Assessment of AI R&D Acceleration at Anthropic in October 2026→Review of the "Risks from automated R&D" section in Anthropic's Risk Report (August 2026)→
• 05Read

  OpenSSH 10.6p2 patches musl sshd as AI-found bug reports surge

  Alpine and other musl systems running sshd without PAM need this critical fix. OpenSSH says AI-assisted reports are surging and the same bugs are being found independently, so the project is now shipping patch releases more often.

  selckin→
• 06Read

  Opposable: dropped links caused 108 of 305 Android agent failures

  Taps and swipes succeeded 99% of the time across 45 phones, so screen understanding was not the weak point. Most failures came from dropped connections, text focus and phones that cannot tell the agent from the owner.

  erikkannike→