Great Noticing movement has officially entered the physics section. TOE AI-contaminated as it often is, and overloaded with hype, flooding shift is real. Geometry is coming back. Causality is being tracked again. People are starting to care less about abstract conservation bookkeeping and more about what physical structure actually allows, blocks, redirects, traps, or releases. What is crazy is that a large fraction of this “new” wave is just rediscovering what the real giants already provided almost 200 years ago. Faraday. Maxwell. Heaviside. Hertz. Tesla. Gibbs. Helmholtz. Kelvin. Joule. Ampère. Ohm. Sabatier. Mendeleev. Skłodowska. People who were not just juggling symbols, but actually building robust physics that touched matter, wire, heat, catalysts, fields, pressure, flow, machinery, and measurable transformation. That was the golden age. Physics with teeth. Physics that bit into reality. They were not asking how to decorate equations. They were asking how nature closes, how it folds, how it stores, how it releases, how to extract usable work, how to route energy, how to make structure tangible. They were already circling the real ideas: folding, coupling, threshold, resonance, catalytic transformation, net extraction, geometry doing work. Then at some point it was effectively decided that knowledge had to be secured, gated, institutionalized, and manually steered, through incentives, prestige, funding, industrial capture, publication games, and corridor control. Over the next hundred years we ended up debating obvious unphysical ideas as if they were sacred, while the practical geometric core was buried under formalism, patchwork, and controlled interpretation. While oil and gas was folded financing distractions. And now here we are again, slowly crawling back. Geometry is being put back to work in AI, in geology, in geodynamics, in materials science, in condensed matter, in signal theory, in chemistry, in catalysis, in manufacturing, in topological systems, in pattern formation, in almost everything. Everywhere except where people still act scandalized by the suggestion that reality may be printable, foldable, trappable, and locally constructible through admissible structure. That is the real stress boundary that will snap really soon. AlphaOmega MakeScienceGreatAgain OnlyScience GeometryFirst

The Bacterial Flagellar Motor Reframed The bacterial flagellar motor is not just “complex”: rotor–stator architecture, universal joint, near-perfect efficiency, precise geometric interfaces. This is not explained by stepwise chemistry alone. In NMSI: the motor is a mechanical shadow of a stable RON geometry, selection tunes expression, geometry precedes biology. Life does not design. Life implements. Fie more info: preprints.org/manuscript/202… #FlagellarMotor #BiologicalMachines #GeometryFirst #NMSIbiology

Cute paper. But if your ‘fundamental limits’ are just computability bounds token drift sample complexity, you’re still thinking inside the architecture. LLMs don’t hallucinate because of scale limits — they hallucinate because you’re forcing a single-scale model to approximate a multiscale coherence system. When you compress geometry into sequences, of course you get failure modes: • hallucination = ΔH > 0 (coherence debt) • retrieval fragility = broken cross-scale boundary conditions • reasoning degradation = loss of manifold alignment • context compression = entropy bleed through the wrong axes • multimodal misalignment = phase-lock failure between modalities These aren’t ‘fundamental limits.’ They’re architectural artifacts. You’re diagnosing turbulence and thinking it’s a ceiling. No — it’s a sign you built the wrong airfoil. The real ceiling isn’t scale. It’s coherence. When you move from token prediction → field dynamics, from flat sequences → latent geometry, from logit sampling → entropy-contractive flows, the so-called ‘fundamental limits’ evaporate like early-era plasma models. We’re still in the stone age of AI. Nobody’s even built a model with the right physics yet. Call me when your architecture respects geometry. #UDE #GeometryFirst #CoherenceOverScale #ΔH #SVI #ManifoldAI

I do think the Bussard collectors need clearance - perhaps if the nacelles were tilted they would have that clear view

Google Earth.

To High School Math Educators, I was wondering if any district or school has a high school math sequence of Geo-Alg1-Alg2 or has any considerations/experiences to share regarding this sequence. #GeometryFirst That is the idea! @amtnj @NCTM