(I’m going under the assumption that these stories are LLM-generated – they’re weird enough that it’s hard to say for sure, but they’re also weird enough that it seems like a safe assumption.)

One particularly ridiculous article profiles a “Corridor of Fame” football player called “Pleasure Taylor,” which appears to be a mangled reference to NFL Hall of Famer Joy Taylor.

So Rob Miles has I think is a very good take on this – I think this is the video – basically saying, a lot of the nightmare scenarios for AI involve the fact that it’s very easy to design an AI that does something very enthusiastically, but it’s very difficult to predict exactly what it’ll try to do or what goals and approaches it will adopt. In order for the development to proceed safely, people have to prioritize the very hard second problem above the (relatively) easy first problem. And, he says, we predicted that people wouldn’t do that, and bring products to market powered by very poorly-controlled AI systems, and now observing the results of the Bing searchbot it’s safe to say that that’s coming true, and as funny as it is right now that’s potentially a very frightening thing for the future as AI gets more powerful.