[noxs]

Everytime when there is some achievement made by Chinese companies, there are always comments like "what about tiananmen square", "I won't trust companies under CCP supervision", "must be fake propaganda", and then being ignorant to the achievement and biased against the not-us-made models, and not focusing on the actual technical part of the achievement itself, like how they can make training so fast and what can we learn from it. Beyond that some US startup CEO even starts spreading rumors like they have 50000 H100 while the paper and model are open sourced since older versions of deepseek that everyone can reproduce and verify it, the theoretical compute need is calculable from model architect.

Sure sure that censorship is a problem, but that's a political background everyone knows, while none of the researchers of deepseek can do much about it, and literally do people think Chinese people like to put more efforts to censor LLM output?

Associate researchers with CCP without any evidence and being ignorant to their achievement is really insulting to deepseek researchers' hardworks.

[elevatedastalt]

It's extra funny coz most of the top American LLM products have censorship stuffed to their gills.

[commandlinefan]

Actually I was thinking I trust this more than I trust OpenAI because I can run it locally.

[creativenolo]

Exactly.

They release the weights so it can be fine tuned to censor/uncensor for your locale and preferences.

[ithkuil]

As with many other words, trust means different things, so you're not wrong

[dutchbookmaker]

It is so boring at this point.

If you want to learn about Tiananmen Square then try reading the book Deng Xiaoping and the Transformation of China by Ezra F. Vogel and Eric Jason Martin.

Amazing book.

Of course, I imagine the people asking this really don't care about Tienanmen Square or China anyway.

It is really someone just being an obnoxious child.

[suraci]

I don't need to read books to know the tiananmen massacre is just as real as the xinjiang genocide

Im mean, there are so many evidences about this, right? unlike gaza genocide which are totally misinformation by russian bots without any credits(I should not talk about things like this, and I think this by my free will)

there are so many photos and videos shows the ccp massacre and genocide

i mean, there is a tank man photo proven this, oh poor guy must be squashed crudely

i don't need to read books, wikipedia and tankman told me everything

it's so terrible and horrible, damn you ccp, I'll expose your sins in every thread about china, by my free will and free speech

[juunpp]

Nicaragua, El Salvador, Honduras, Indonesia, Vietnam, ...

[suraci]

your thoughts containing errors, must be greyed out

[juunpp]

What do you mean? Are you too stupid to read some literature?

[sdesol]

> Sure sure that censorship is a problem, but that's a political background everyone knows, while none of the researchers of deepseek can do much about it, and literally do people think Chinese people like to put more efforts to censor LLM output?

I think it is a knee-jerk reaction without understanding how LLMs work. The beauty of all of this is, we can use DeepSeek and still give the CCP the middle finger. I don't know why people don't realize we can easily add a layer above DeepSeek to never ask it for political/historical information and we can easily create services to do this.

We should be celebrating what is happening as this might force OpenAI and Anthropic to lower prices. DeepSeek is FAR from perfect and it would be stupid to not continue relying on other models, and if DeepSeek can force a price change, I'm all for it.

[Barrin92]

There's a famous quip from the 2008 US election cycle, where Biden remarked that every sentence Rudy Giuliani forms consists of "a noun, a verb and 9/11".

This is pretty much the same thing on a national scale. US discourse in particular is increasingly petty, bully-like, disrespectful, ignorant or straight up hostile as seen with the tone concerning Indian tech workers recently. Even Latin Americans or Europeans aren't safe from it any more. I'm afraid we're only at the start of this rather than the end as China and others catch up or even lead in some domains.

[eunos]

That is a quite good psyop since everybody only focussed on muhh censorship rather than pondering technically.

[code_for_monkey]

people have been scare mongered about china for so long theyve just absorbed the racism into their bones, half this comment section is just "tiannemen square!" or "winnie the pooh!"

[EGG_CREAM]

Stop it with the false equivalency. China’s government is not Chinese people. You can’t be racist against a government.

[suraci]

i like the 'hate the government but not the people' here

[osti]

If you think ppl on HN or elsewhere aren't racist against the actual Chinese people, you are just incredibly naive.

People have been talking about the Chinese like automotons of the government with no agency of their own for a long time now. However it's the same for all of humanity. In China in the Mao era, the slogan was to free the Western capitalist society from repression. It's the same old talking about enemy camp and assigning no free will to the people.

All this is to say, people here don't think of the Chinese as equals. That is the real core of racism, not about saying something against the protected race of the day.

[noxs]

Many people hide their racist views to Chinese people under the guise of criticizing the Chinese government.

[neochief]

If I'd be Chinese government, I'd love people spreading the idea like yours, though.

[buyucu]

Americans are having a crisis of confidence as their technological lead is crumbling before their eyes. Expect more tantrums of this kind.

[vunderba]

I kind of agree with you.

I mean shocker: large language model trained in mainland China may have censorship around topics considered politically sensitive by Chinese government, more news at 11. Can we move on?

But it's also an easy low-hanging fruit if you want to add a comment to a Hacker News Post that you otherwise don't know anything about.

[xgstation]

nit: replace "may have" with "has"

but yes I am tired of seeing this kind of "news". they don't carry much useful information, more like noise nowadays

[benterix]

The point is, the amount of mutilation done to models released by OpenAI and co. is enormous. So I very much hoped a Chinese model would be much more free from all this kind of BS. But when I think deeper about it, they actually had no choice: imagine the amount of criticism they would face. At this point, the only accusation you can hear from their competition is "For sure they used more housepower" which seems to be quite weak.