OpenAI’s page is titled “Accelerating engineering for the physical world,” but the material available does not explain the project or provide evidence.
The Latest
AI And The Push To Accelerate Engineering Beyond The Screen
How Z.ai’s Legal AI Earns A Near-Top Benchmark Rank With A 1-In-5 Pass Rate
Z.ai says GLM-5.3-FlashX-Legal scored 19.2% on Harvey’s all-criteria benchmark; Vals’ archived leaderboard places it seventh of 81 models.
GPT-6 Astra Tried A New Kind Of AI Repair: Editing Another Model
Researchers report that GPT-6 Astra edited two Qwen3.5-0.8B weight tensors, raising its MATH-500 score while reducing—but not eliminating—loops.
A Closer Look At CWE-bench V2’s Expanded AI Coverage
Collinear AI has released 100 public security-fix tasks across 91 CWE categories; its leading reported model scored 67.8% on pass@1.
Why Claude Code’s Faster Message Threads Won’t Change Your Token Bill
A report says Claude Code has added message threads to speed up work, while token costs stay the same. The feature and speed gains are not detailed.
Launch HN: Momentic (YC W24) – Agents That Catch Bugs By Using Your App
Momentic, a YC W24 company, says its AI testing platform runs end-to-end app checks and helps teams verify code changes before release.
When Almost Everyone Farmed: Understanding Work In The AI Era
A historical account traces agriculture’s shrinking workforce, from 19th-century farms to today, and examines what the shift can—and cannot—say about AI.
What A Garment Wage Board Could Mean For Bangladesh’s Economy
Bangladesh appointed a six-member garment wage board after protests over a 30,000-taka wage demand disrupted 529 factories, according to BGMEA.
Anthropic’s Claude Adds A Trail To Text It Helps Create
The Independent reports that Claude will leave a trail indicating AI involvement, but the method, scope and timing are not available.
From AI Data Centers To Compute Campuses: Who Covers The Costs?
OpenAI confirmed it may call data centers “compute campuses,” as grid costs, voluntary industry pledges and water use remain under scrutiny.
