Posts
All the articles I've posted.
-
From Sand to Stargate: The CPU and GPU Chip Supply Chain, Silicon to Silicon Wafers and Beyond
An end-to-end research survey of the CPU and GPU chip supply chain in 2026 — from a single Appalachian quartz mine and Ukrainian neon plants up through 300mm wafers, EUV scanners, leading-edge foundries, HBM/DRAM/NAND, advanced packaging, design and EDA, ATE, the geopolitics of CHIPS Acts and export controls, and the hyperscaler demand wave that is now consuming most of it. With 130+ primary references — SEC filings, government reports, SEMI/SEMI Europe data, Yole/TrendForce/SemiAnalysis/CSIS, and company disclosures.
-
Navigating Blind: How Drones Fly Without GPS — and Why It Matters More Than Ever
GPS jamming and spoofing are now routine in contested airspace, and indoor/underground operations have always lived without it. This post is a technical survey of GPS-denied drone navigation in 2026 — VIO, visual SLAM, terrain-relative navigation, satellite image matching, magnetic anomaly navigation (MagNav), LiDAR SLAM, and multi-modal factor-graph fusion. Deep dives into satellite image matching (the most promising daytime approach) and MagNav (the dark horse that works at night, in fog, and is unjammable). With lessons from Ukraine, the open problems, and a practical stack for builders.
-
The Edge of Mathematics Research: A Citation-Grounded Survey of Open Problems and Recent Breakthroughs (2019–2026)
A subfield-by-subfield survey of the mathematics research frontier, grounded in specific papers from Annals, Inventiones, JAMS, Acta, Publications of the IHÉS, and the work of recent Fields medalists. Covers number theory, algebraic geometry, topology, geometric analysis, PDE, probability, combinatorics, logic, representation theory, harmonic analysis, dynamics, mathematical physics, and the rise of formalization. The through-line: mathematics in the 2020s is fusing — Langlands now spans number theory, geometry, and physics; condensed and prismatic methods unify analysis and arithmetic; and Lean is starting to absorb the load-bearing proofs.
-
The Edge of Artificial Intelligence Research: A Citation-Grounded Survey from ICML, ICLR, NeurIPS (2023–2026)
A domain-by-domain survey of the artificial intelligence research frontier, grounded in specific papers from ICML, ICLR, NeurIPS, and adjacent venues (CoRL, CVPR, Nature). Covers natural language processing, speech, video, sound, robotics/VLA, biology, 3D generation, diffusion architecture, and the renaissance of reinforcement learning and world models. The through-line: pretraining is no longer the frontier — test-time compute, generative simulators, and embodied grounding are.