I just had a thought about that law they passed in the US to prohibit the distribution of 'intimate'(why did they use that language instead of obscene like most legislation does?) deepfakes.
At first I was pleased it outlawed distribution of deepfake porn instead of possession, as possession would mostly lead to prosecution of randos who put the wrong combination of words in a box.
But I was just thinking(like literally right as I was reading this post): what if the whole purpose was to force cloud providers to put monitoring infrastructure in place that will ultimately be used for other purposes(either by the feds or the woke mafia)?
I mean most AI specific providers (openAI, grok) already had that, so no worries there, but the law might make even GPU compute providers like runpod, etc, put in monitoring infrastructure.
Kind of surprised I didn't think of it earlier, my tin foil hat is normally pretty tight.
The code demanded will likely be terribly designed, and ignored by the actual program. Just have it return a 0 no matter what, and you have done your duty.
That came as a response to two things: chatgpt taking the #2 spot to only fans for subscribers and grok desecrating protected holobunga survivoooooors with bikinis. (Anne frank getting plowed by German shepherds didn't help either but this of course was done with locally run and trained SD models, of which they have no control over)
Intimite deepfakes law is a predictable result of this :
“Kimi2.5 is open source and performs as well as Anthropic’s claude.
What normally happens when rich companies have no protection from competition is - mass lobbying for “safetyism” rules that free companies cannot comply with.
(See Facebook’s Nick Clegg hired to do just that)
step1- expect to see articles from our state media : “anthropic keeps your kids safe from transing themselves”
step2- mandatory ID for anyone who uses AI to “keep kids safe” (it will be largely old people’s fears of youth)
step3- Only government supervised AI - to keep kids safe”
I just had a thought about that law they passed in the US to prohibit the distribution of 'intimate'(why did they use that language instead of obscene like most legislation does?) deepfakes.
At first I was pleased it outlawed distribution of deepfake porn instead of possession, as possession would mostly lead to prosecution of randos who put the wrong combination of words in a box.
But I was just thinking(like literally right as I was reading this post): what if the whole purpose was to force cloud providers to put monitoring infrastructure in place that will ultimately be used for other purposes(either by the feds or the woke mafia)?
I mean most AI specific providers (openAI, grok) already had that, so no worries there, but the law might make even GPU compute providers like runpod, etc, put in monitoring infrastructure.
Kind of surprised I didn't think of it earlier, my tin foil hat is normally pretty tight.
The code demanded will likely be terribly designed, and ignored by the actual program. Just have it return a 0 no matter what, and you have done your duty.
Sounds like some shit my coding agent would whip up.
Exactly. The real question is what information has to go through it, and will it be stored for review?
That came as a response to two things: chatgpt taking the #2 spot to only fans for subscribers and grok desecrating protected holobunga survivoooooors with bikinis. (Anne frank getting plowed by German shepherds didn't help either but this of course was done with locally run and trained SD models, of which they have no control over)
Intimite deepfakes law is a predictable result of this :
“Kimi2.5 is open source and performs as well as Anthropic’s claude. What normally happens when rich companies have no protection from competition is - mass lobbying for “safetyism” rules that free companies cannot comply with. (See Facebook’s Nick Clegg hired to do just that)
step1- expect to see articles from our state media : “anthropic keeps your kids safe from transing themselves”
step2- mandatory ID for anyone who uses AI to “keep kids safe” (it will be largely old people’s fears of youth)
step3- Only government supervised AI - to keep kids safe”