The Brief
New briefs daily around 7 AM Eastern
Thinking Machines releases Inkling 2 Mini, an open 26B laptop model
Artificial Analysis scores it 20 on its Intelligence Index, level with 27B-class open models, at $0.59 per task against $3.53. That puts a mid-size open model on a MacBook under Apache 2.0 for about a sixth of the cost per task.
- 01Read
Claude-built Lean proof claims the Generalized Riemann Hypothesis
Terence Tao and Scott Aaronson both say verification is still pending. The claimed proof is a 17.4M-line Lean artifact built with Claude, so its acceptance rests on machine checking and on whether the formal statement matches the theorem.
- 02Read
LangSmith data puts gpt-6-luna on 27% of production teams
It climbed from #30 to #1 since its September 22 launch. Claude Sonnet 5.5 and GPT-6.1 Sol each reached 22% of teams but handle under 2% of calls, so teams are trying new models widely while keeping most traffic on a cheaper default.
- 03Read
MiniMax model breaks NIST candidates and recovers 94 McEliece keys
The attacks are executable and the code is public, so they can be checked rather than taken on trust. A model producing working breaks of submitted cipher designs moves AI-driven cryptanalysis from a claim to something security reviewers now have to plan for.
- 04Try
AI SDK 7.0.178 closes a replay bypass in tool approvals
Before the fix, a client that controlled message history could replay a policy-denied tool call as approved and have it run. Apps using experimental_toolApprovalSecret relied on that signature as their human-in-the-loop guard.
- 05Full story
Signal blocks screenshots to keep phone AI assistants out of chats
Assistants such as Gemini Intelligence and Siri Onscreen Awareness read decrypted messages straight off the screen, outside the end-to-end encryption Signal provides. Signal now blocks them by default, with an opt-out.
A dated brief from the vibe-coding frontier. Today’s Intel.