How a chip is made, why the leading edge sits with one company in one place, and how the chip war between Washington and Beijing has been fought, from the US-Japan precedent of the 1980s to the 2026 truce. Follows the AI concentration from Nvidia through TSMC, ASML, and the memory makers, whose 2026 boom made memory half the industry. Opens with a dated table of the state of the art and a Who's Who.
Collection 6 — Technology
The industries at the frontier and the chains beneath them.
What a large language model does, how it is trained, why capability has risen on a predictable curve, and what the industry looks like in October 2026: the frontier labs heading for the stock market, a build-out of about $800 billion a year, agents that work for hours, the open-versus-closed split with China, the harms and the laws, and whether the money is ahead of the technology. Opens with a dated table of the state of the art and a Who's Who.
The machine that moves: sixty years of the factory and the failure at the kitchen, the AI turn that lets a robot learn instead of being programmed, the five million robots in factories and the million in Amazon's warehouses, Waymo's half a million paid rides a week, the humanoids, the cheap drone's remaking of war, and the arguments over jobs and China. Three worked examples: a humanoid's hardware costs, a factory line's payback in Ohio and Shenzhen, and a robotaxi's costs per ride. Opens with a dated table of the state of the art and a Who's Who.
The physics one level deeper than the headlines and the ways to build a qubit; error correction, now shown below threshold on three platforms, and the roadmaps to fault-tolerant machines around 2029; the algorithms that matter and the advantage claims, weighed; the threat to encryption and the argument over Q-Day; the sensors and clocks already in service; and the companies, from Quantinuum's listing to Washington's new stakes in quantum firms. Worked examples include a surface code's overhead, one bank's migration, and IonQ's revenue against its market value. Opens with a dated table of the state of the art and a Who's Who.
How information travels, who owns the networks that carry it, and who controls them, as of October 2026: the internet's layers and its history from the telegraph to the smartphone; packets, addresses, routing, and encryption; the fiber backbone, the exchanges where networks meet, and who pays whom; the submarine cables that carry nearly all traffic between continents, and the cuts in the Red Sea and the Baltic; broadband at home and what it costs; mobile networks, the airwaves, and the auctions that sell them; Starlink and its rivals; the cloud, its data centers, and the outages and drone strikes of 2025-26; the 2.2 billion people still offline; who governs the internet; and the hacks, record shutdowns, and national walls that threaten it. Where the field argues, each side's case comes before the course's reading. Worked examples include a household's month of data, the half-second before a web page appears, a broadband bill, and a video call from New York to London. Opens with a table of key numbers, a Who's Who, and a list of terms.
How processors work, who makes them, and why architecture and software decide who wins, as of October 2026: what a processor does and why free speed-ups ended around 2005; making one core fast; memory, the wall every processor hits, and the 2026 shortage that made phones, PCs, and consoles dearer; x86, Arm, and RISC-V, and why the instruction set now decides who may build a chip; how the GPU went from pixels to AI; the chips in phones, PCs, consoles, and cars; server CPUs; Nvidia's accelerators, their challengers, and the clouds' own chips; the networks that join hundreds of thousands of GPUs, racks heading toward 600 kilowatts, and CUDA's hold on the software; how computing is measured and priced; and the companies and the money. Where the field argues, each side's case comes before the course's reading. Worked examples include why a chatbot's speed is set by memory, what Arm earns from a chip, the cost of a GPU-hour, and where a gigawatt's dollars go. Opens with a dated table of the state of the art, a Who's Who, and a list of terms.
How attacks on computers and networks happen, what they cost, and how households and organizations prevent them, find them, and recover from them, as of October 2026: risk and the NIST framework's six functions; who attacks and how an intrusion unfolds; passwords, multifactor authentication, and passkeys; software flaws and patching; scams and fraud, from business email to the scam compounds; ransomware and extortion; suppliers, open-source software, and the cloud; water, power, and hospitals; what states do in cyberspace, from the Iran war to China's Typhoon campaigns; AI on both sides; detection, response, and the reporting deadlines; backups and rebuilding; cyber insurance and the war exclusion; the law in America, Europe, and Britain; and privacy and data brokers. Where the field argues, each side's case comes before the course's reading. Two worked examples run through it, a household whose accounts are taken over and a manufacturer recovering from ransomware. Opens with a dated table of key measures, the state of the art, a Who's Who, and a list of terms.** **