Scientific Fraud Is The Most Baffling Thing Ever To Me Like Do They Think They're Just Going To Make

Scientific fraud is the most baffling thing ever to me like do they think they're just going to make a huge breakthrough and no one will notice that it's fake by trying to replicate their results

I've described trying to steer Stable Diffusion as "trying to drive a car by talking to it." It's really great at things like, "hey car, get me to starbucks." Because it knows how to get to starbucks and that's not really conceptually complicated. But if I'm like, hey car, drive towards starbucks for five minutes, but then do a hard u-turn and then pop a wheelie and spin around for a little bit, and then drop by Mike's place, and also get me to McDonalds, it gets really confused. Because individually, all of those things are things it can do, but together, you're basically trying to wrestle this machine into submission and control. The reason AI image generation is handy for a lot of Slop use-cases for a lot of people is because they just want an Image. they don't really care what. And it's really good at that. But the more granular you get, the more it gets confused.

maisy mouse library moments

Fluffy bunny tummy by Sparrowl on Twitter

I do consider Wolverine's interactions with Jean and Scott in the first X-Men movie to be sort of an inverse unicorn hunter scenario. sometimes the unicorn hunts you. aggressively.