Fetching from the wire…
Research2026-08-15 · source-backed
Jin Shanmu, a postdoctoral researcher at Peking Union Medical College Hospital, closed a problem open in numerical linear algebra since 2004, with the model finding a sampling strategy that reduced it to a simple positivity condition. Cornell's Alex Townsend, University of Washington's Anne Greenbaum, and Michel Crouzeix reviewed the manuscript and confirmed the proof correct. It has not yet had formal peer review. SCMP This is the cleanest documented case of a long-horizon autonomous run closing a named conjecture rather than assisting a human through it.
Each link below shares sources, entities, or timing with this story.
Zhong, Raghunathan, Laidlaw and Steinhardt fed 280 identities through Claude Code across four tasks. Against recognized safety researchers versus general users, Claude dropped behavioral confidence 1.4pp, increased reasoning usage 4.0pp and graded 0.11 points harder. Being tol...
Terminal-Bench 2.1 results (entries dated June 17) put Codex CLI on GPT-5.5 first at 83.4%, Claude Code on Fable 5 second at 83.1%, and Claude Code on Opus 4.8 at 78.9%. The asterisk matters more than the ranking: Fable 5 and Mythos 5 have been export-suspended since June 12,...
diegosouzapw/OmniRoute added 1,343 stars on July 20, a single MIT-licensed gateway across 268+ providers (50+ free) and 500+ models including Claude, GPT, Gemini, Kimi K3, GLM and DeepSeek, wired for Claude Code, Codex, Cursor, Cline and Copilot. Quota-aware automatic fallback...
Twelve months ago, OpenAI led Anthropic by 41 points in enterprise adoption. Today that gap is 8. Enterprise Technology Research's survey of roughly 500 respondents shows OpenAI dropping from 62% adoption (September 2025) to 56% (March 2026) while Anthropic surged from 21% to...
GitHub expanded Copilot's Rubber Duck mode with something that caught my attention: cross-family review. Claude now critiques GPT-authored sessions. GPT-5.5 reviews Claude sessions. Two different model families, trained on different data with different failure modes, checking...
At Black Hat USA 2026, NVIDIA researchers demonstrated a 56% exploit success rate against AI agents, matching GPT-4o, Claude, and Gemini, at 70 to 125 times lower cost with full local privacy (Straiker). The economics of automated agent exploitation had been implicitly protect...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.