The Structure Carries More

Researchers built a staple-fiber core with two continuous helical filaments, compared physical yarns with a finite-element model and varied filament radius, elastic modulus and winding pitch. The publisher-deposited abstract reports that pitch had the largest effect on negative Poisson behavior, followed by radius and modulus; its optimized combination—0.05 mm radius, 85 GPa modulus and 4.3 mm pitch—produced a minimum Poisson ratio of −2.54. The accessible record does not expose specimen count, wash or fatigue testing, garment integration, comfort or production yield.

Five common-form silicone modules use different inextensible-fiber patterns to produce extension, expansion, twist, shear and bend. The authors combined finite-element simulation, UV-marker measurements and a shear-and-extension Cartesian platform that traced 4 × 4 mm square paths for ten cycles per plane. Testing used one specimen per module under quasi-static, open-loop and unloaded conditions; shear–axial coupling reached 47.3%, while payload, force, extended fatigue, closed-loop control and manufacturing reproducibility were not evaluated.

A wearable sweat-lactate system combines silane-network-encapsulated lactate oxidase with a polyvinyl-chloride diffusion layer, then adds pH and temperature calibration plus iontophoretic sweat induction and microfluidic collection. The publisher-deposited abstract reports a 0–60 mM range and 82 paired samples from 33 critically ill patients: marginal R² was 0.840 for sweat against blood lactate and 0.811 for the integrated module against arterial blood gas, with AUROC of 0.943 for hyperlactatemia and 0.785 for mortality. The small observational study does not establish interchangeability with blood testing, outcome benefit or independent prognostic value.

Google began rolling post-quantum cryptography into Google Meet for laptop web clients and meeting-room hardware on October 12, with mobile support described as coming soon. The update covers Workspace, Workspace Individual and personal Google accounts, requires no administrator or user action and is designed to create no visible change in a call; Google allows up to 15 days for rollout. The first-party post does not name the exact handshake or algorithms, quantify overhead, describe fallback behavior or confirm completion, and it offers no evidence of a present quantum attack.

Meta introduced Muse Wonder with Active Memory across conversations and Active Focus, which tries alternatives, asks clarifying questions and checks before important actions; the model is available through Meta's Model API and Muse Code, while the larger Wonder Full is limited to Muse Code subscribers. Meta's model card says its MemoryBench and FocusBench are internal, comparison harnesses are mixed, visual and spatial reasoning remain limited and multi-step errors can compound; it recommends human confirmation for irreversible actions. Independent evidence for real-world reliability, privacy and long-running task safety is not yet established, and Meta says a fuller safety report will follow.