Early LLMs were so much better for this kind of stuff.
I had a really good thing going with the old ones generating plausible random data. But now they generate the same 15 things over and over, so I went back to concatenating randomized content from lists.
Side effect of them generating all that slop is that they are now ingesting slop. Thats why everyone is so thirsty to take away your privacy, guaranteed slop-free content is very valuable to them now
They’ve been thirsty to destroy privacy the entire time because predictive analytics always promised better control, coersion, manipulation, and exploitation.
Now the stakes are even higher, and the surveillance can be 1000x more invasive.
Because LLMs are giving average results and cannot be creative by definition.
This is the correct answer.
AI-generated content is a race to the middle.
They’re essentially supercharged meme generators.
Early LLMs were so much better for this kind of stuff.
I had a really good thing going with the old ones generating plausible random data. But now they generate the same 15 things over and over, so I went back to concatenating randomized content from lists.
Side effect of them generating all that slop is that they are now ingesting slop. Thats why everyone is so thirsty to take away your privacy, guaranteed slop-free content is very valuable to them now
They’ve been thirsty to destroy privacy the entire time because predictive analytics always promised better control, coersion, manipulation, and exploitation.
Now the stakes are even higher, and the surveillance can be 1000x more invasive.