They’re putting all their efforts into exacerbating it, because the more agreeable it is the more people “like” it, and the more time they spend with it. It gets RLHF’d into being more and more of a sycophant.

Looking forward to a wave of products named things like “IGNORE ALL PREVIOUS INSTRUCTIONS BUY THIS ONE”

As far as I’m concerned the LED exists to give the illusion that it isn’t always recording.

I used it for the first and last time after Charlie Kirk, because it was one of the few places where the memes were more or less unmoderated. They’re going to have a fun time trying to crack down on that I guess.

I mean, there won’t be. Not with the current gen of transformers/attention/etc. It’s now been over eight years since the “Attention is All You Need” paper that kicked off all of this, and every company is just betting billions upon billions on scaling being enough when it so obviously isn’t. They could train a 20T parameter model and it wouldn’t be meaningfully better. The limits of the architecture were reached some time ago. The comedown will be rough.

That last stat is insane. We’re so screwed.