The models in this group describe technology from the outside: how it is adopted, why the less elegant option often wins, how platforms change once they dominate, and what might happen if systems become more capable than their makers. Several are about artificial intelligence, and they are among the most argued-over entries in the ledger.
How technologies spread
Moore's Crossing the Chasm locates the hardest step in adoption between enthusiastic early adopters and the pragmatic majority, who want proven solutions rather than new ones. Gartner's Hype Cycle describes expectations rising, collapsing and then settling. Richard Gabriel's Worse Is Better explains why a simpler, less complete design that spreads early can beat a better one that arrives later, which is his account of Unix and C defeating more elegant rivals. Japan's Galapagos-ka describes the opposite fate: products so tuned to one isolated market that they cannot compete anywhere else.
What platforms become
Cory Doctorow's Enshittification describes a sequence in platform businesses: good to users to attract them, then worse to users to serve business customers, then worse to both to extract value for shareholders. Kentaro Toyama's Law of Amplification offers a broader frame, drawn from development work in India: technology amplifies whatever human capacity and intent already exist, so it widens gaps between institutions that work and those that do not.
Agreement and scale
Two classic puzzles from distributed computing set limits on coordination. The Two Generals Problem shows that no finite exchange of messages over an unreliable channel can guarantee agreement, and the Byzantine Generals Problem shows that consensus fails once a third or more of participants may lie. The Universal Scalability Law adds that adding workers eventually lowers throughput, because the cost of keeping them coordinated grows faster than their contribution.
Machines more capable than their makers
The AI entries range from fiction to formal argument. Asimov's Frankenstein Complex names the old fear that a creation will turn on its creator. Bostrom's Paperclip Maximizer, the Orthogonality Thesis and Instrumental Convergence argue that intelligence does not imply sensible goals, and that almost any goal pursued capably leads to acquiring resources and resisting shutdown. Sutton's Bitter Lesson observes that general methods which scale with computation have repeatedly beaten methods built on human expertise.
Where it applies
The adoption models are used in product strategy and investment, where the chasm and the hype cycle shape launch plans and valuations. Worse is better is a warning to anyone polishing a product while a cruder rival ships. The distributed-systems puzzles are the practical foundation of databases, payments and blockchains. The AI models now inform policy and laboratory safety practice, whatever one thinks of their conclusions.
Where it is argued over
Several popular models rest on less evidence than their popularity suggests. Researchers who have tested the hype cycle against the histories of actual technologies have found that many do not follow it. Enshittification is a polemical description whose economics, two-sided markets and lock-in, are well studied, though its inevitability is disputed. The AI arguments are the most contested: critics argue that the orthogonality and convergence theses describe idealised agents rather than systems that actually exist, and Kurzweil's Law of Accelerating Returns depends on which curves are chosen for plotting. The Blub Paradox, Graham's claim that programmers cannot see the power of languages above their own, is an argument from experience rather than a tested claim.
Why it is worth knowing
Technology is usually discussed as a set of products. These models treat it as a set of processes: adoption, lock-in, decay and scaling, each with its own logic. That view makes it easier to predict what a new technology will do to the organisations that adopt it, and to weigh the largest claims being made about AI with the same scrutiny the ledger applies everywhere else.
The entries behind this essay: Model in Technology & Software. Every subject has its essays on the essays page.