I found the comparison between open web standards and laws interesting because both depend on people recognizing some shared rules even when enforcement isn't always immediate. It makes me wonder whether the openness of the early web depended partly on a level of trust that doesn't work as well when this decade's biggest players have enormous financial incentives to ignore those norms. If so, the problem might not just be that companies are violating individual rules, but that the social conditions that made those rules effective have changed. Once one company breaks the rules, the rest now have a competitor who did it before them that they can point to and feel validated by. New players in the market have to break the same rules to compete; it’s just terrifying that nobody knows how invasive that cycle could potentially get.

I think the educational consequence here is easy to overlook compared with the more obvious concern about AI-generated code overwhelming maintainers. Open-source communities don't just produce software; they also give inexperienced people a way to learn by participating in something real and receiving feedback from more experienced developers. If AI-generated contributions make communities more restrictive, beginners could lose one of the few places where they can gain experience without already having professional credentials. This makes the issue feel less like a problem affecting programmers and more like a problem affecting how people enter an entire field.

There is an interesting paradox here: AI can make information easier to access in the short term while potentially making the internet harder to access in the long term. If creators respond to AI scraping by putting their work behind paywalls, the amount of high-quality information available publicly decreases. When AI systems increasingly rely on existing material, humans have less incentive to create new material openly. I hadn't really considered the possibility that the biggest consequence of AI scraping might not be the content it takes, but the content people eventually stop making available.