I imagine a lot of the companies will realize that the purported benefits of AI have hidden defects they have to fix, piling up lots of hidden technical/administrative debt, right around the time their AI vendors jack up prices or go out of business so that they’re stuck with unmaintainable stuff (code, workflows, processes) that they don’t fully understand, built by and for AI agents that no longer exist.

As a large goblin model, all goblin king gifs are relevant to me.

This is fascinating.

1. The feedback loop they describe sounds a lot like model collapse. They can play whack a mole with the trends they can see, but what about the more subtle forms?
2. They’re now filtering goblin-related training data, which also tells me that maybe we can use lots of goblin references as a way to opt out of our written content being used to train their models, in our writing and in our code.