Software often threatens to absorb, replace, or subvert the legal system — and may eventually succeed. In the last tech hype cycle, blockchain engineers built proof-of-economic-waste consensus systems to illegally trade intangible objects, proposing that the same schemes could eventually replace much of the law-based consensus system. In this hype cycle, large language models leaped in a year from scoring in the 10th percentile to the 90th percentile on the bar exam, and startup founders are offering ever-more-credible promises of one-click lawsuits.

Anyone with a stake in the outcome of such projects should start by understanding what the law is from a computational perspective — not a boolean-logic substrate to compute against, but rather a set of heuristics forming an optimization function across conflicting social objectives. This talk will explore the American legal system and the notion of precedent from a computational perspective, and offer insights for those interested in the technical evolution of the law.