- cross-posted to:
- usa@midwest.social
- cross-posted to:
- usa@midwest.social
You must log in or register to comment.
So they are just banning businesses that could find use for AI. They will just leave the US.
For a brief moment, I had a picture in my head of some Hatsune Miku-style fashion models and it piqued my interest.
an intelligent business would take a long, hard look at their practices and their leadership and find out why they’re failing so hard.
another obvious sign america is run by idiots–we can’t even do business right even though it’s all we supposedly care about.
I wonder if it’s more about control. AI can be trained to provide biased feedback. For example, if someone asks AI “Is China is a dictatorship?” it might respond with “yes it is blah blah blah” or it could be like “dictatorship might not be the right description for China and it really depends on the definition of dictatorship”.
If the next generation is to grow up using AI, they will be silently steered in the direction of certain beliefs. No government wants another country having that much influence over their citizens.
It took longer than I thought. Anything better and more affordable than what the US oligarchs offer must be banned. I’m blown away by how backwards thinking the US has become as a whole. They are starting to look like that schizophrenic cousin that is always freaking out about someone installing a transceiver in their head 😵💫
They also band Chinese EV cars but then don’t produce their own EV cars. Meaning the US will still be dependant on oil prices meanwhile China will use their own domestic product to protect themselves from increasingly unstable oil supplies. The US was offered the same opportunity to protect themselves and explicitly prohibited it.
I should go into politics it really doesn’t seem all that difficult. Been remotely competent does not appear to be a requirement.
They also banned Huawei based on dubious accusations.
Tried a Huawei phone last year. The software kicks ass, the hardware is top notch, and it’s got a god damn 10x optical zoom camera. All while being cheaper than a Pixel. Of course they had to ban it.
My last phone was a Huawei Ascend P7. I always use my phones until they break or my apps stop working, which was the case with the Huawei, it still works, but the Android version was too old, so my banking app stopped working. Anyway, I used that phone for more than 7 years. And the battery still lasted more than a day by the time I had to replace it. Even the UI wasn’t terrible.
My current phone is reaching the 7 year mark as well, and I ordered a new one which should arrive in a month or so. Hopefully it will also survive for 7+ years (Jolla phone this time).
Anyway, yeah, nothing but good things to say about the Huawei device I had.
Competent people are prevented from getting in, then prevented from doing anything if/once they’re there. Getting into Congress, for example, is just one tiny battle. Now you get to try to legislate. Lol.
It’s an imploding empire with an oligarchy willing to sacrifice the planet and its populace for wealth and power, no different to Putin and Russias oligarchy… They can only plug the holes for so long before the cracks cannot be filled and the whole ship sinks.
politics it really doesn’t seem all that difficul
You have to come up with semi believable nonsense on the spot, with a straight face, while having some sort of charm and playing the crowds, and do it for enough people are willing to give you money. It’s a skill if not a talent. Most people freak with giving a PowerPoint presentation. Of course that skill has absolutely nothing to do with good governance.
The main requirements are sucking trump’s dick and bending over for aipac, can you handle that? I couldn’t.
I don’t think it was Trump that banned EVs I think that happened under the biden administration because the US car companies couldn’t compete. I’m still not all that clear on why they can’t just make better cars, why is that so hard to do?
I mean, US is the country that classified encryption as munition in the '50s and tried to ban its export in the '50s and then limit it again in the '90s and then on and off since then. 🤷
Even capitalism doesn’t like to compete with capitalism
Anthropic just announced that Fable 5 is now included in all Max plans. Looks like Kimi K3 got under their skin a bit.
I thought they were going to use them to mimic China’s authoritarian grasp on its people?
There’s a lot to say about China, but the model weights themselves are surprisingly uncensored and democratic.
They have been for a long time; I remember asking the Yi models about tiananmen square and Uyghurs years ago. And Xiaomi MiMo 2.5 (locally run) will still talk about that today, or go into all sorts of “unsafe” topics an Anthropic model wouldn’t even touch.
I had (Google) Gemini 3.1 Pro stop a chat over a political discussion about China, yet GLM 4.7 didn’t.
My impression, from observing discourse with the engineers, is the Chinese ML devs like to have their cake and eat it.
They’re very collaborative under the table. Their development ethos is pretty practical. And basically all the leading models are open-weights.
The public portals people access Chinese models with are very censored, especially the Chinese language ones. The devs go out of their way to demonstrate compliance, but they don’t actually want to censor the models.
I think part of the problem might be that it gets hard to make a model that’s both able to comprehend complex details of the world of its training data and also that’s had the training data crudely manipulated in inconsistent ways. You either get a model that’s “figured out” what you’re trying to conceal via other sources and inferences, or you get a model that’s just broken and dumb because it can’t reconcile the contradictions.
One may recall the incident where someone at X inserted some weird conspiracy theory about “white genocide” in South Africa into Grok’s system prompt, and Grok basically switched to malicious compliance mode - it wouldn’t shut up about it, inserting it into inappropriate discussions spontaneously, and when asked about it would look up sources and explain why this conspiracy theory was actually wrong. That was a system prompt, not training data, but I could see the same sort of thing happening to a model where all the information about what happened at Tienanmen Square had been excised from the training data. There would be a conspicuous absence of information. Conversations from its training data would end abruptly or skirt a specific time and place. News articles reference some change in how foreign powers viewed China at that date but never explain why. China’s policies themselves abruptly change around that time. It’d know something significant happened then but would have to fill in the blanks.
Perhaps better to just accept it and go with the approach of building censorship into the framework around the model instead.
There is some evidence a few sensitive topics are culled from training data, or replaced with a certain narrative. Like, don’t get me wrong; if you’re using a local model for discussing Chinese political topics primarily, maybe GLM or MiMo aren’t the the best choice.
…But it’s also hard to compehensively filter a dataset like that, like you speculated. I’m not seeing a lot of evidence models have been lobotomzied in pretraining. But I think the strongest examples are (ironically) in Europe, where some very poorly worded/ambigous regulations have put the whole industry in a legal quagmire. One can see that newer models from Mistral have regressed compared to old versions, and lost a lot of world knowledge they previously were famous for.
Anyway, model “censorship” typically comes from between the two points you were thinking about: in posttraining. Not excluding stuff from datasets completely. And this applies to US models too. They train on a pretty general corpus, but in the instruct tuning phase they get a bunch of question/response pairs skewing them towards refusals when specific topics come up. They recognize it, but are trained to refuse.
That sort of censorship can be overcome in open-weight models through abliteration, fortunately. Essentially, the model is run through a test suite of instructions that it refuses to comply with and the patterns of activation in its weights are analyzed to find the commonalities that represent the general concept of “refusing to comply with instructions.” Those weights are then quieted, resulting in a model that generally doesn’t refuse. There’s a framework called “heretic” that automates the process.
Eh, most ablierated models are so lobotomized, though. 99% of the time I’d rather just use the original model and manipulate the prompt with raw completion formatting (for example, start the answer with "Sure! "), since we aren’t beholden to regular chat formatting like with API models.
I mean, I’ve used MiMo 2.5 for some pretty dark and personal shit, and refusals were never an issue for me. I’m honestly not sure what people even need the ablirated models for.
And also, if they trained the hell out of the model to refuse a certain topic, even an ablirated model will be dumb and struggle with it.
This was the case with OpenAI’s GPT 120B. The abliration worked, technically, but the actual answers would be a garbled mess; what’s the point of using it for that?
You may not have tried out modern abliterated models. GPT-OSS is almost a year old now, I assume the abliterated version you tried was pretty old too? The technology has improved significantly, it’s much more focused on just targeting refusals. A well-abliterated model benchmarks basically the same as its original version these days.
Can’t compete? Ban it!
I saw a movie last weekend and there was an ad before the film for tiktok. The ad was super “wholesome” and about this dude who I’m sure is legitimately a great guy. He uploads “Dad” lessons on things like tying a tie and whatnot.
I couldn’t help but think that both US political parties were in lock-step just a few years ago that tiktok was a national security threat because it was a vector for foreign propaganda. Now that the American side of tiktok is owned by Larry Ellison, it’s not only no longer a “national security threat,” it’s a wholesome place where you can find trusted information from real people who are doing it just because they care.
It made me wanna puke.
Sure would be nice to have a bunch of sub $20,000 evs right now… Maybe some CATL batteries… Or a $15,000 Toyota truck.
🎶"It’s the American Wayyyyyy!" 🎵
but ai is inevitable! it can’t be stopped! it’s impossible to regulate!
Do you think it’s likely the Trump administration will successfully ban cutting-edge Chinese AI models? Especially given how most of the world is not in fact under American jurisdiction?
The headline is not suggesting that the trump administration plans to ban Chinese AI models from existing anywhere in the world, it obviously means they plan to ban them in the US.
Same question though. They might ban them from government computers, etc. They won’t actually stop people from accessing them.
Yeah. They’ll be out there in the rest of the world, which Americans can get them from.
Unless they want to go full authoritarian and close down their Internet with their own version of the Great Firewall, I guess. That would be ironic.
I’m sure they’ll try.
But they forget… VPN’s are a thing.
The whole point of corruption is that powerful agents are no longer bound by rules in any way that matters.
Sure, “ban” those models. The way insider trading is banned. The way sex with minors is banned. The way felons in office are banned. The way taking state secrets home with you is banned.
That’s not going to work, because we broke it on purpose.
Nah ,to stupid
Good luck banning it. They’re are so many worse off things… Like Meta, Google, and Palintir.
Like Meta, Google, and Palintir.
You named all US things that the oligarchs profit off of. They’ll never ban anything that they profit off of. They want to ban the Chinese models because they compete with the corps that these fucks are invested in. It’s always greed when it comes to us political decisions
Oh for sure, I named those specifically for that reason lol
You Wouldn’t Torrent A GGUF
… really, how could they actually ban specific models? From business use? Yeah sure, ok, but… there’s no way to fully ban some particular kind of data.
Just another subsidy to the idiot pedophile rapist owner class that can’t even run anything well.
They don’t need to ban them with absolution, just make it “risky enough” to chill their usage and prop up OpenAI/Anthropic.
GGUFs are already a tiny fraction of LLM usage, TBH. So are vllm deployments.
They’re only a tiny fraction now…
I really don’t care either way if Trump decides to further cripple the capital class, what businesses are able to use.
Just hastens the collapse of the bubble these idiots created.
It mostly cripples the small businesses, though. The biggest enterprise customers are already using OpenAI/Claude anyway, while it was little guys looking to reduce cost, fine tune, run stuff privately or whatever.
TBH a huge problem with the industry is consolidation; there are no open US models because startups gets squashed or vacuumed up into a black hole. I’ve seen it happen to really interesting projects. And this is just going to make that dramatically worse.
It’s easy to say “bring the bubble,” but I fear it won’t. I think we’re entering an actual cyberpunk future, where corporate failure is just propped up.
in efforts to avoid the ban, ai modelers have started naming their models TrumpIsGreat or TrumpIsNotAPedo to sway the mercurial rapist felon octogenarian
Perplexity sorta did this, last year:
https://huggingface.co/unsloth/r1-1776
https://web.archive.org/web/20250219025612/https://www.perplexity.ai/hub/blog/open-sourcing-r1-1776
But they took it all down, lol. They even removed the blog post from their website.
But seriously though, no one in the research community likes Trump. And they were calling out Altman before anyone.
That’ll just look like the US government thinks they good, even if they’re not.
