Archive · June 2026
Neural Radiance Fields & 3D Gaussians: Your Building, Fitted Not Modelled
How neural radiance fields and 3D Gaussian Splatting turn a phone walk-around into a metric-accurate building capture — the foundation, the maths, and a Hack.
Voronoi & Delaunay: One Point Set, Two Trades
How Voronoi cells and their Delaunay dual turn one seed field into both a panelisation and a framing logic — with a runnable d3-delaunay browser hack.
SCRIPT Wants Humanoids to Obey Language — Read the Provenance First
SCRIPT couples language, state, and action in one diffusion transformer for humanoid control — but its wins are self-reported sim results. What AEC should verify.
Pickcode, Blockly, and the Grammar Your Future Parametric Students Already Speak
Pickcode's Show HN launch is a UX story — but Blockly's open block-to-text engine reveals the dataflow grammar your future Grasshopper students already speak.
The Building That Clusters Itself: A Cancer-Triage Algorithm Lands on the Smart Estate
A new arXiv variational Dirichlet process method speeds up clinical subtyping — and gives building-OS designers a way to let data choose its zones.
Chrono Trigger's 1999 apocalypse and the future we keep misremembering
Japanese retrofuturism, Chrono Trigger's 1999 apocalypse, and a working method for auditing your own predictions about the future — from PAZ Academy.
The dice my hands roll: how diffusion policies learned to vote on a move
KeyStone and DSSP both fight the brittleness of diffusion-based robot action policies — judge-free consensus sampling and full-history conditioning, explained.
Voronoi: the pattern a body uses to spend the least material
A PAZ foundation essay on Voronoi diagrams — the geometry of nearest-neighbour packing, its Delaunay dual, and how it zones CLT, lattice infill and façades.
Make Your Own Website: The Plain-HTML Stack That Still Opens in 2070
A hands-on PAZ tutorial: build a website from one hand-written index.html with VSCodium — the durable, vendor-free stack that still opens decades from now.
The self-certifying cache: why LAWS could make on-site robot AI provable
LAWS proposes an inference cache with a deployment-time error bound you can check without ground truth — what it means for on-site robots and BIM AI tools.
