Out of curiosity, Carolina Figueiredo noticed that three apparently unrelated quantum field theories all forbid exactly the same set of particle collisions. The shared structure turned out to be curves drawn on surfaces — the framework now called surfaceology, where Feynman diagrams thicken into 2D surfaces and spacetime stops looking fundamental. Extending it to the full Standard Model is open, but Figueiredo thinks it's reachable.

Two newly-confirmed phases — ice XXI and ice XXII — have unit cells of 152 and 304 water molecules, the most structurally complex any team has measured, both produced under diamond-anvil pressure. A separate Lausanne group reports a 'plastic ice VII' that appears when high-pressure ice is heated to around 500°C. A 2018 simulation cataloged 75,000 candidate phases; experiment is finally catching up.

Reports of Chinese labs siphoning capabilities from frontier APIs are real, but Lambert argues the 'distillation attack' framing conflates jailbreaking and API abuse with a routine, legitimate technique used across the industry. The policy risk is non-trivial: regulation written in panic could ban open weights without touching the actual abuse it targets, and would mostly punish the academic and startup ecosystems still doing distillation honestly.

How ASML — a Philips spinoff that competitors initially mocked for outsourcing components — bet on tin-droplet plasma EUV and a dual-stage immersion architecture, then waited thirty years for the bet to pay. Hacker takes the physics seriously: 13.5 nm light from laser-vaporized tin, mirrors that would be flat to a few atoms if scaled to the size of Germany. The strategic story (Intel, TSMC and Samsung buying a combined 23% of the company in 2012 to fund commercialization) tracks the technical one closely.