Comment by ballon_monkey
7 days ago
The 2 things people need to remember:
1) China can (and does) use the models to influence the west. They train in false information about Taiwan and Hong Kong. Or pretend like history is in favor of China.
2) Ignoring the models containing false information, they are incredible. But you should be scared of running inference via the model creators directly. If you think your data is safe compared to running it via model providers in the US ( either frontier or model hosts like fireworks.ai ) then please let me know your bank details so I can poke around.
https://thereallo.dev/blog/claude-code-prompt-steganography
Why should I trust a US company more than a Chinese one?
Your exhibit is an obvious anti-distillation technique. You can expect far worse from less-regulated companies.
Neither Anthropic nor OpenAI are regulated in the "regulated business" meaning.
2 replies →
Agree, China believes in Climate change and are at least taking steps to address it. Personally this is making me trust them more than the USA because, facts.
I mean lesser of two evils thinking, if one is intentionally leading us towards climate disaster, while the other isn't then yeah. What else can be said? Should I trust the authoritarian country who believes in engineering and science, or the one that doesn't?
>Agree, China believes in Climate change and are at least taking steps to address it.
China brought 80GW of new coal power online last year. The US added 0, and plans to add 3GW next year (we all know why).
China doesn't care about climate change, they care about energy independence, and conveniently have very little natural fossil fuels besides coal. Which they heavily mine and utilize.
10 replies →
China emits 3x more co2 than the US and last year their rate of increase was 2.5x that of the US.
https://www.worldometers.info/co2-emissions/co2-emissions-by...
49 replies →
USA and Europe are basically equal in terms of yearly production of renewable energy. Texas + California alone rank 6th in the world.
>China believes in Climate change and are at least taking steps to address it
I've got a bridge to sell you!
Exactly my question
You shouldn't trust either but also this is whataboutism.
No, it is not. Whataboutism is a specific term created (by Cold War propagandists) to refer to when the US criticizes the USSR about "X," and the Soviet Union replies by asking about "Y" in the US.
It is not a term created to refer to when some American criticizes any of America's official enemies for "X," and someone else reminds you that America also does exactly "X."
10 replies →
No it's not. There is a difference between a comparison and whataboutism
There's nothing intrinsically wrong with "whataboutism" and it is in fact often illuminating.
1 reply →
[flagged]
Been there. Great place. High standard of living, everyone is very productive. Still a dictatorship, and you can't speak ill of the government, but a successful one.
3 replies →
What if they don't live in USA?
Abstracted from the HN specific audience, there’s around 17% chance they already do, which is to compare to the "around 4%" chance they live in USA.
https://www.worldometers.info/world-population/population-by...
This is such an hilarious answer. For most of the non-Western world, we don't distrust China because they have created benefits for the developing world. Whereas the USA has been outright imperialist for the past 60-70 years, undermining and overthrowing developing democracies, putting in dictators, using international organisations as debt traps, waging and funding wars that put our lives and economies at risk.
It is so pathetic how American's see themselves, and are so deeply afraid of China. I am far more fearful of the predatory nature of the USA and its agencies than I could ever be of China who would never have any interest in me.
4 replies →
[flagged]
1 reply →
The ROI for (white) men going to China for a period of their lives is extremely high. Where else are you going to get your “Abg cmo”
1) I'm not using AI to bicker over fringe political shibboleths.
2) I don't think it is any more or less safe to put my code on a Chinese server versus an American one. A Chinese provider also isn't liable to spy on me for the feds, as OpenAI and Anthropic certainly do.
Different county different feds both spying I'm sure.
Sure but if you are not Chinese or in China, then the chances of negative consequences to you from the Chinese feds is vanishingly small due to lack of ability to do anything that affects you.
Meanwhile if you are in the US, DHS has already subpoenaed social media sites looking for people who made anti-ICE posts and I can't imagine they consider subpoenaing AI conversations off limits https://www.nytimes.com/2026/02/13/technology/dhs-anti-ice-s...
3 replies →
Regarding point 2, I don't trust my data being safe running inference on model creators api, but neither do I trust US providers. Both use it for their own benefit, the only difference is the country of origin. The US has a lot more legal safeguards for this but I don't trust they don't do it regardless.
legal safeguards only make sense only when they’re enforced. With how “move fast and break things” Silicon Valley is law is always playing catch up (at your expense)
It’s perfectly fine for the “West” to influence the world though, right? Or is it only a problem because… they’re Chinese?
Why would it be a problem that they're Chinese? It's a problem because their country is ruled by an authoritarian regime. The West, by contrast, doesn't have a singular arbiter of truth, it has a plurality of perspectives (as much as certain interest groups would really rather that was not the case). It's not perfect, but those of us who have memories of living under authoritarian regimes can tell you there's no comparison.
> It's a problem because their country is ruled by an authoritarian regime
Hey, have you seeing what trump does with your (presumably) country? Maybe wars? Maybe market manipulation? Mayde pedo right covering on the government level? Maybe bubbles and threats to EU? What an ignorance. You live with old stories, not the current state of the world...
8 replies →
If being ruled by an authoritarian regime is the problem, then wouldn't both Chinese and American models be problematic?
European models too, if they had any.
4 replies →
The United States has propped up some pretty horrific, murderous, bloodthirsty regimes over the past century to ensure they maintain a proxy presence across the globe. If supporting or engaging in an "authoritarian" regime is your litmus test, then the United States fails resoundingly.
3 replies →
West has though a history of supporting regimes far more hideous than Chinese, and ready to push narrative whenever possible. It is well known idea BTW in political thought that democracies are far more reliant on propaganda for control than authoritarian,which due to their strong grip on political discourse do not need high quality propaganda anyway.
TLDR: American propaganda is not any better than Chinese, neither have the best interests of my country in their minds.
Have you got examples of GPT/Claude/Grok influencing people?
I think there's no denying they each have an ideological bent (as is their right as private companies). I have had Copilot deny me access to historical information on ethical grounds, even though I don't think anyone would find it controversial (clearly it was overtuned, GPT and Claude had no problem answering the same question). What is different though is that they are each allowed to have their own perspective instead of a singular mandated one.
6 replies →
1. https://www.uea.ac.uk/about/news/article/fresh-evidence-of-c... -- (2023) Fresh evidence of ChatGPTs political bias revealed by comprehensive new study
2. https://www.ox.ac.uk/news/2026-01-20-new-study-finds-chatgpt... -- (2026) New study finds that ChatGPT amplifies global inequalities
3. https://stevepavlina.com/blog/2026/03/chatgpts-political-bia... -- (2026) ChatGPT’s Political Bias
4. https://pmc.ncbi.nlm.nih.gov/articles/PMC10623051/ -- (2023) Revisiting the political biases of ChatGPT
5. https://www.theverge.com/2024/2/21/24079371/google-ai-gemini... -- (2024) Google apologizes for ‘missing the mark’ after Gemini generated racially diverse Nazis
https://i.imgur.com/3zWdJpz.png
vs
https://i.imgur.com/ta6g1d2.png
A major thing I noticed is that trying to get ChatGPT to say something negative about Sam Altman is like pulling teeth.
Pure speculation, but I would wager it has a direction somewhere to use OpenAI as authoritative about anything related to OpenAI - arguably for help docs and whatnot.
But the impact does stay the same.
Why did you mention Grok, because it undermines your argument.
https://www.theguardian.com/technology/2025/may/16/elon-musk...
https://www.trtworld.com/article/24cbdb873a6b
On (1), models are an aggregation of large volumes of data sources, whatever the culture producing them, you'll get the average bias of that culture.
We see that on what minorities are associated with inside the model, or how things that aren't online will have a completely different weight. Or how 2/4/5/8ch or X will be disproportionately present in specific models despite being the places where facts go to die.
I see your point, but this isn't exactly true, it only takes a single person to bias a model by deleting specific training data or over training on certain facts.
But hasn’t it been the same with US since the WWII? US has influenced the world through various ways sometimes even weaponising human rights to spread American superiority and its narrative?
While true, you neglect the fact that US corporations have some autonomy from the government. I doubt that any of the current American AI corporations get regular dossiers of desirable responses to sensitive political topics from Washington and implement them. That would be a reason for a major lawsuit and a major scandal.
In China, the image of the country and the preferred narrative is under much tighter control of the government and local AI shops won't have any autonomy in this regard at all. Either obey or get shut down.
BTW What you mean by "weaponizing human rights", exactly? I am curious. If anything, I would say that the US foreign policy didn't promote human rights sufficiently, especially in Latin America, where the "bastard, but our bastard" attitude was typical.
OTOH in Europe, US human rights policy was probably relevant in saving some dissenters in the former Eastern Bloc from torture or execution.
> you neglect the fact that US corporations have some autonomy from the government
I'm not sure if this is correct. It seems like every big US corporation is changing their policy based on the views of the administration in place. As an example during Biden's time the DEI was in full motion in every corporation and in current administration it's the other way around. Also I remember as soon as Biden won the election twitter and meta suspended the profile of Trump. So I hardly think that there is much room for autonomy. Also the latest export bans on Antropic and OpenAI models kind of make your argument weak, of course the company can sue the givernment, but the national security comes above all.
> BTW What you mean by "weaponizing human rights", exactly?
US has been using the violations of human rights to impose sanctions or to especially get countries in line which are not supporting the US agenda when it comes to global politics. However, they were more than happy turn a blind eye if the respective country that violates human rights (most middle eastern countries) if they are allies of US doctrine.
21 replies →
>US corporations have some autonomy from the government
This is correct...because they just buy the government when they need to.
In China the state rules the corporations and in the US the corporations rule the state.
Neither is a system worth fighting for. Swiss style democracy maybe but not autocracy and not oligarchy.
>OTOH in Europe, US human rights policy was probably relevant in saving some dissenters in the former Eastern Bloc from torture or execution.
Yeah, a bit like how Russia saved Edward Snowden.
The US will use human rights as a club to beat its imperial rivals with but when it has deemed torture and arbitrary execution in the interests of its imperial power it has adopted them enthusiastically.
When US allies use these tools and worse they are excused.
There is basically nothing which US rivals do which the US wouldn't also do under similar circumstances.
1 reply →
> I doubt that any of the current American AI corporations get regular dossiers of desirable responses to sensitive political topics from Washington and implement them.
That's completely true. What happens is that you get hit with a wall of bureaucratic threats and nonsense by the government out of nowhere, and usually for some unrelated matter. It's like a high level version of getting pulled over for not using your turn signal. Which is why if your company has good legal, they often act in a highly proactive manner about that kind of thing.
What constitutes sufficient rabble rousing to get put on the government's shitlist differs between both countries, but there are plenty of things that will send American politicians into having a conniption, and it's not a static criteria. McCarthyism is a pretty easy example. I'm no Muskboy, but we can point to the procedural harassment he recieved over hitting a blunt on the Joe Rogan podcast to be a related matter. Criticizing the genocide happening in the Levant is frequently saber rattled by politicians as something they're interested in making prosecutable, and there are examples of institutional authority targeting and attempting to punish people over it.
2 replies →
> you neglect the fact that US corporations have some autonomy from the government
In theory.
In practice, Trump picks up the phone and say "jump" and the CEO on the other end says "how high ?".
And if you say no, well, we saw what happened when Anthropic said no.
12 replies →
but you are not using chinese models to learn about taiwan, you use chinese models to do everything BUT learning about taiwan, so no issue there
> China can (and does) use the models to influence the west
OK now that's false information.
You can uncensor, tweak or fine-tune open-weight models, but not so easy on a proprietary model from some cloud provider.
It's probably true (why wouldn't it be?) but the missing context is that the US does it even more.
1. Get the free Chinese model. 2. Jailbreak it 3. ??? 4. Profit?
> They train in false information about Taiwan and Hong Kong. Or pretend like history is in favor of China.
American labs can open their models or their old models at any point if they actually care about this, but until the day OG GPT 4 isn't averrable to download or Claude 3 then they're only pretending to care about this because they can profit from restrictions.
It's incredible how accurate this is towards US models:
> The 2 things people need to remember:
> 1) USA can (and does) use the models to influence the rest of the planet, and put political pressure. They train on stolen data, hide information, gate keep, who knows what they're hiding. In favor of the USA, nevertheless.
> 2) Ignoring the models containing false information, they are incredible. But you should be scared of running inference via the model creators directly. If you think your data is safe compared to running it via model providers in China ( either frontier or model hosts like z.ai ) then please let me know your bank details so I can poke around.
There's an irony in it. I have many reasons to believe China has its own geopolitical narratives and goals just like the US does. If someone dislikes when the US does it, why ignore when China does the same thing and praise them for their efforts? These models exist and sure, they accomplished something interesting. Use them with caution but I don't appreciate the political narratives about good and bad guys and AI models and politics.
> 1) China can (and does) use the models to influence the west. They train in false information about Taiwan and Hong Kong. Or pretend like history is in favor of China.
One of the interesting things is that through a fairly rudimentary process which is being done by 3rd party amateurs who've downloaded the open models, models like Qwen 3.6 35B-A3B (or 27B) can be fully 'uncensored' when turned into GGUF files.
I have an uncensored Q8 version of Qwen 3.6 35B-A3B here that will very happily output information about Tiananmen Square, Uyghurs, human rights in China, or indeed can even be instructed to write an intentionally absurd vitriolic screed against the CCP. The same uncensored 27B (dense) will do the same, just at a slower token/s rate.
Similarly there's 'uncensored' variants of Gemma4 31B and other western trained models, which once put through the same process, will also discuss or write just about anything you want, bypassing whatever internal guard rails were attempted in the training data set.
edit: more concerning, and a very legit concern, is that a model is only as good as the sum total of its training dataset, so if something is trained on a steady diet of news sources like Peoples Daily, Xinhuanet and similar in the English language, then it'll have a greater percentage of CCP-approved media publications in its training dataset. No amount of uncensoring it will help with that after the fact.
Never forget that those Chinese tanks did a sick burnout on tank man's corpse and then t-bagged the remains.
Propaganda exists everywhere and it's your duty as a citizen in a democracy to inform yourself and properly evaluate bias in the media you consume, such as "Kill the boer" by South African mus
I have seen some uncensored versions of existing open models made by 3rd parties, I wonder how they actually work and in what way are the outputs different.
Somehow chineese make less troubles and more good to the world than americans at this point... Something something about ai benefiting all humanity, something something ai being open and stuff, like OpenAI.
Chineese simply delivering what americans promised.
Tell me: why is EU safe from Trump forcing AI companies to cut access to EU?
> Tell me: why is EU safe from Trump forcing AI companies to cut access to EU?
Well, I think they might end up doing it to themselves by imposing regulations that US companies are unwilling to put up with.
It's so naive to think that US companies won't do (or aren't already doing) the exact same thing but for money.
Or even similar ideological goals... (See: grok)
> pretend like history is in favor of China.
Are you saying that history has a verdict, and it disfavors particular 3000 year old cultures?
> 3,000 year-old cultures
Like Western culture?
Sure? Someone calling a model biased for "pretending history is favorable to western culture" would sound pretty provocative.
So you agree - it would be stupid to say "pretend like history is in favor of X civilization"
You think that the US based models aren't doing the same?
Is anyone using these models as their daily chat driver? I am skeptical.
The vast majority of Westerners will interact with ChatGPT/Claude/Google in a browser. They'll use these models to try to save some money coding.
All models/data sources are biased - you need to understand inherent biases in any data source.
The first point is a strawman - these models are not going to be used to set foreign policy on Taiwan - it's to write code etc.
Likewise risk of data exfiltration and misuse isn't model specific. Indeed it's not even the biggest data risk - there are much larger risks from the data we know companies like Meta and Google already collect.
Bottom line - a model you can download and run on your own private infrastructure is always going to be safer than anything accessed over the internet - as in that case it's not even just the hosting service that's the issue.
i'm also influenced by claude every single day with its constant preaching about certain values.
But Dario said (and maybe more ppl) that Chinese models are just distillation of their model and training data, so I guess your first point is invalid?
> But Dario said (and maybe more ppl) that Chinese models are just distillation
"But Dario said" ... yawn.
I am increasingly convinced that "they distilled us" is as much US FUD as "it was made by communists". Especially since its mostly the US tech-bros who are coming out with that tiny violin.
People telling me the Chinese models are distilled just because it says "I am Claude" when asked is also lame.
I am not the only one, look at this post on interconnects about Kimi K3 for example:[1]
[1] https://www.interconnects.ai/p/kimi-k3-the-open-weights-esca...
It probably was distilled but the point is why is that an issue at all, Claude is just distilled from all our work. Why is stealing our work fine but stealing Anthropic's work bad.
Anyone doing so should have zero guilt because it was already stolen goods to begin with.
"Or pretend like history is in favor of China."
What does this even mean?
Like it will tell you the PRC (current China) was right to go to civil war with ROC (former China, current Taiwan)
Personally I'm more worried about a US provider/model censoring for BS reasons than Chinese ones. Biology 101 is too dangerous for Anthropic. If I want an AI to scrutinize the CCP for Tiananmen Square, Taiwan or Tibet because I'm bored, I'll use a US or heretic model.
Your data isn't safe from exfiltration no matter who the host is. You should host the models locally if your data is truly sensitive.
Regarding 1.
We’re literally tearing down monuments to slavery and civil rights.
Religion is now determining law in much of the nation. Having a miscarriage? Good luck since politicians have decided their God doesn’t want you to have access to basic healthcare.
The government has defacto control over domestic LLMs.
Let’s worry about our own historical record.
> China can (and does) use the models to influence the west. They train in false information about Taiwan and Hong Kong. Or pretend like history is in favor of China.
I am not Chinese and I'm not defending the Chinese, but I see this argument come up a lot.
In practical terms it is US-sponsored FUD.
Why ?
Because the hard reality is that what you say is simply not going to affect 99.9999999999% of users.
Is it realistically going to affect anyone using an LLM in coding ? No.
Is it realistically going to affect anyone using an LLM in $anything_else_not_politically_sensitive ? No.
Does anyone seriously use LLMs for researching politically sensitive matters ? No.
Just as there is plenty of information out there on the US's less than perfect history, there is also plenty of information out there on the various Chinese politically sensitive matters. You do not need a Chinese LLM to find out about it, all you need is a search engine.
People absolutely use LLMs to research politically sensitive topics.
I wish they wouldn’t, but people use LLMs as their general search engines now.
> People absolutely use LLMs to research politically sensitive topics.
Note I used the word "seriously", I meant it in its fullest form, i.e. serious people.
I'm not interested in what-if arguments based on "you can't fix stupid".
Stupid people also blindly believe whatever a US LLM tells them without any form of verification, hallucinations and all.
Most people on this planet would agree that a Chinese LLM is perfectly usable for all tasks except asking about politically sensitive matters.
And for most people on the planet, that is just fine. They can get their political information elsewhere.
3 replies →