Pip: The Hedge has been calling things early since 2008, and timothymccandless is keeping that tradition alive with a look at what happens when the legal system meets a technology it genuinely wasn't built for.
Mara: This episode is about copyright law under pressure from AI — the cases in court, the doctrine that's breaking, and the four scenarios for how it might resolve. Let's start with the reckoning itself.
The Copyright Reckoning: How AI Rewrites Everything — Including the Law
Pip: The central tension here is structural, not procedural. Copyright law was built on two assumptions — that expression is scarce and that copying is detectable — and AI has quietly demolished both without anyone agreeing on what replaces them.
Mara: The post frames the active litigation — the NYT suit against OpenAI, the Authors Guild actions, Getty Images versus Stability AI — and lands on this: "Fair use was designed for humans doing creative work. An AI processing 100 billion tokens of human writing to produce commercial output doesn't fit that mold — and courts know it."
Pip: Which means the doctrine isn't just strained — it's pointed at the wrong subject entirely. Fair use assumed a person with expressive intent on the other end. That assumption is gone, and courts now have to either stretch the framework until it's unrecognizable or admit it simply doesn't apply.
Mara: The market-harm prong is where it gets most concrete. The four-factor fair use test has always weighted market harm heavily, so if AI output replaces demand for the original work, the transformative-use defense takes serious damage regardless of how technically different the output is.
Pip: And then there's what the post calls the rewrite problem — which is the sharper edge. If AI can take any copyrighted work and produce a cleaner, updated version of the same ideas, copyright only ever protected the specific expression anyway. AI just industrializes the paraphrase at a scale that makes that distinction feel hollow.
Mara: Four resolution scenarios are on the table. Licensing regimes modeled on ASCAP and BMI are called the most likely near-term outcome. Output rights carved out separately from training rights come next. Congressional action is flagged as least likely given how slowly IP law moves. And fair use expanding until enforcement atrophies is described as unlikely but not impossible.
Pip: The honest bottom line, as the post puts it, is that copyright was a bargain — temporary monopoly rights in exchange for eventual public domain contribution. AI broke that bargain in both directions.
Mara: Creators will get something. AI companies will pay something. Neither amount will feel adequate. That's the pattern, and the post doesn't pretend otherwise.
Pip: The legal system will patch something together — it just won't be intellectually coherent. That's a fair description of most major technological transitions and their aftermath.
Mara: The pressure is real and it's building. Worth watching which of those four scenarios starts hardening into precedent first.
The Hedge 