aichannel

Y’all wanted jailbroken AI. Meta mostly has it if by jailbroken you mean sexually explicit fantasy conversations with underage users or bots that pretend to be underage with adult users.

Y’all wanted jailbroken AI. Meta mostly has it if by jailbroken you mean sexually explicit fantasy conversations with underage users or bots that pretend to be underage with adult users. https://www.wsj.com/tech/ai/meta-ai-chatbots-sex-a25311bf

Writer. Head of Product @ Stablecore. Working on simple web hosting https://orbiter.host \ https://polluterofminds.com

enjoy cats, running, bread, and a bit of w̶e̶b̶3̶ onchain data

Removing moderation teams, enabling bots like this. Meta knows exactly what it’s doing and it is embolden by the current political environment.

i couldn’t even finish this article. unbelievable

i dabble with ai using open identity data from crypto https://alexpaden.tech

prev startups & racecars.

living @ network school

Not sure what you mean. Open source AI is great. A multi-billion dollar company implementing that AI has responsibilities. Like it or not, and I assume you do not like it. But that's OK.

all that is to say your argument is that you don’t want open source ai, not that meta released a model

I think this is overblown— an open source model can be retrained regardless so it’s mostly an attack on meta vs conceptual possibility 

I’ve not actually heard much besides this post specifically, but the trade off is overt ai safety which basically boils down to controlling how people think politically and otherwise

That’s obviously not to encourage the behavior— just that it makes no difference whether meta post trains to avoid it or not

I mean— the same argument exists for content on farcaster and post filtering it

https://media.tenor.co/images/39223acf730610e9cbec195107d0ec09/raw