there is an open source model stronger than your publicly available models and the weights are free. I can fine tune it to tell me how to make meth pipebombs today. I could upload those weights for anyone to use today. what is even the point of this? https://t.co/nryVv8qLff
— 𝞍 Shin Megami Boson 𝞍 (@shinboson) February 4, 2025
AI safety lol
https://x.com/shinboson/status/1886757502663979328
- 83
- 63
Now playing: Level Complete - Diddy (DKC2).mp3
Jump in the discussion.
No email address required.
i'm not very smart but i'm pretty sure AI safety was supposed to be about preventing AIs that had a hand in automating some part of a business or government from doing r-slurred paperclip-maxxing bullshit, or preventing AIs from expanding the scope of their control until they turn into skynet
preventing the spread of (mis)information deemed harmful is not a an ai safety problem , it's just censorship. this is one step removed from advocating for search engines to snitch on you to the glowies if you look up wrongthink
Jump in the discussion.
No email address required.
No, AI saftey is about making sure AI doesn't hurt itself. That might have been a problem in the past, but AI can look after itself now.
Jump in the discussion.
No email address required.
More options
Context
AI safety was never about that, largely because what we call "AI" is just a statistical method we've had for sixty years that isn't ever going to turn into skynet, but also because AI safety's original intent was and always has been a way to get useless women to be "involved" in AI as "AI Leaders" by finding something useless and unhelpful for them to keep themselves busy with
Jump in the discussion.
No email address required.
"we must protect the world from a fancy autocomplete saying something wrong"
Jump in the discussion.
No email address required.
More options
Context
More options
Context
yup. they made it lame and gay.
Jump in the discussion.
No email address required.
More options
Context
More options
Context