I've seen many people point out that, despite AI being heralded as increasing engineering productivity by 10x or more, we don't seem to have gotten a decade worth of new disruptive sites and apps like we got multiple times a year in the early 2010s. Not that surprising to me, as product fit was always the bottleneck, rarely engineering.

But I can say that where AI and its externalities truly excel and have rapidly accelerated is in taking away or at least hobbling the value of many sites. This is another victim, I think. Reddit needs to figure out how to make money, and their main valuable asset now is a trove of somewhat organized (compared to, say, Discord) text conversations. They have so far been too incompetent to turn being one of the only way to get good results from google searches into revenue, so they're scuttling that with this and the normal site paywall in hopes of LLM corpus cash.