16383 Followers
335 Following
Senior Editor, Tech & Policy, The Verge
When LLMs go off the rails, people often try to find out what happened by asking the bot itself. But they should be skeptical of the answers. As @alexhanna.bsky.social explains, the only real way to find out is to demand transparency from the creators. www.theverge.com/x-ai/758595/...
LLMs make things up — including about themselves.
"is speech a sandwich-" [i am immediately tackled and dragged off the stage]
darkman is worth at least 1/3rd of the mcu
yeah unfortunately nearly all of these rule a lot more than the results of hollywood interpreting things correctly in the '10s
I'm less comfortable with legal crackdowns on bot speech than many people, but social opprobrium seems like fair game.
I keep wondering why I've become so much less reticent about condemning AI chatbot fails than social media moderation fails, and I think I've come down on: overmoderating social media risks removing genuinely valuable human communication, overmoderating companion chatbots risks mostly reducing slop.
constantly thinking about this
IT'S ME BOY I'M YOUR PS5 SPEAKING TO YOU INSIDE YOUR BRAIN
i'm like the cooler for hope
the problem is there already are a bunch of guys on that, they just want to replace capitalism with masculinist feudal monarchy
I'm so confused, she had NotebookLM... recap recaps, which she had to look up herself? How much time could this possibly have saved.
