For the military, I’m sure this is a cost-cutting measure. I’m guessing they get to fire some people And make the rest of them work double time, but with AI assisting them.
For the working people, it’s another argument that you can point to to get people radicalized :D
Yeah they all do that to each other. Pretty standard practice actually.
The big deal is the low cost, and using weak GPUs, to train a comparable AI.
Trumps gonna be Trump though. Embarassing to live here.
Yeah, unfortunately we’ve known that for a while.
There were actually some Google employees that protested this, and they’ve all been fired, of course.
Exactly. The people who spout “apolitical” all the time are just a part of the global elite.
It’s the rich that will tell the poor to stop being political and go find a job.
It’s the white people who will tell the black people to just stop doing crime instead of getting political.
It’s always your boss that says to keep politics out of the office, just show up on time and do your job.
The oppressor always tell the oppressed to stop getting political. Because getting “political” is the only way that you could potentially stop the oppression.
Bro, just one more compromise with the fascists, bro. Bro, if we just compromise one more time, everything will be fine!
A note about those queues.
I’ve been thinking a lot recently about the difference between policy and material conditions. The USSR wouldn’t have had queues (or very long queues) if they were a rich country like the United States. Any socialist country that is rich would have plenty of materials that nobody would ever have to wait in line (at least not a long line, no longer than in current day USA)
Of course, under socialism, certain excesses would be limited, but at the same time, if your country literally already had what it needed because of the overcapacity built by capitalism, then under socialism, basically all needs would be met very easily.
I need to do some more reading by other people on this topic because I’m no expert, but it makes sense to me.
Libs are worried about it, reference: https://analyticsindiamag.com/ai-news-updates/deepseek-crushes-openai-o1-with-an-mit-licensed-model-developers-are-losing-it/
So my experience right now with using & reading about AI and the opinions and feelings from people about the AI is that China is still a little bit behind, but catching up so insanely fast that everybody should just agree that they’re already ahead of the United States and that the race is on. (Oh and they’re running on last year’s hardware because they’re banned from getting the latest Nvidia chips, so they’re actually training better AI, with weaker processors, and less data. Though they are probably leveraging ChatGPT for some data generation, which is what most people who build open models do anyway.)
Eric Schmidt was a previous CEO of Google, and still a board member. 9 Months ago: https://www.reddit.com/r/singularity/comments/1cnce06/eric_schmidt_says_the_us_is_23_years_ahead_of/ Last month: https://thehill.com/policy/technology/5042031-artificial-intelligence-danger-eric-schmidt/
Eric is about as libed up as you can get (Co-wrote a book with Henry Kissinger…) So if he admits that China has caught up and will surpass us, it’s because the evidence is OVERWHELMING. This would be the equivalent of orthodox Jews admitting that Israel is committing a genocide. If THEY are saying it, it’s been true for a LOOONG time and is undeniable.
Also, they are open sourcing their models, which is freaking awesome. Anyone can use the absolute massive beast of R1 and compete with o1 directly, without even training their own models.
Also Colloquially, AI’s trained in China absolute dominate the open source rankings. Qwen2.5, DeepSeek 3, DeepSeek R1, QwQ (Qwens equivalent to R1) and derivatives from these lead almost every category. LLama (Facebook) is right behind though. For small models, Phi (Microsoft) is up there (Good for running on your phone or laptop, but not the smartest).
AI moves so fast that by the time you’re done reading this something will be out of date 🙃
What I’m curious about, and have not yet asked on 小红书, is whether there is an online space that kids aren’t allowed in so that people can talk about things that they don’t want kids to be exposed to, but isn’t illegal.
LGBTQ topics come up first. Their excuse is, hey, we don’t want kids seeing that, but it’s totally fine if you are and it’s not illegal. So I wonder if there’s a place that you can talk about that stuff that is age restricted.
On a separate note, to keep everything in one comment, I do think that America brains have some issues with censorship that they don’t understand. Sure, the government doesn’t have that many rules around censorship, but every single platform does, and you’re gonna get kicked off or banned from one platform or another for saying certain things. It’s just that in America, those things are inconsistent, and you can get banned for whatever reason the CEO wants you to be. This just fractures our online environments and creates tribalism, which leads to more infighting, not less.
Don’t feel bad about feeling like you’re going down a conspiracy theory rabbit hole. Obviously be conscious of that and be careful, but the fact of the matter is that a lot of anti-China sentiment is a conspiracy theory in the first place.
I am a white guy, not an Asian American, but I’m glad that Asian Americans like you have class consciousness and understand what this country is truly about with regards to its propaganda. Obviously, most people who have immigrated here from communist countries have done so because they are capitalist friendly. I tried explaining this to a friend of mine when we were talking about the USSR. He said, oh, all the people here say how bad the USSR is. And I said, of course they do. They fled the USSR. They obviously don’t agree with its practices. The vast majority of people that stayed behind actually have positive feelings about the USSR In retrospect.
To be fair, I probably pushed him a little hard when I said that a lot of people that fled were the slave owners and horrible people, and he felt that I was implying that the descendants of those people are all horrible people too, which I didn’t mean to say. But that’s another off-topic rant.
One of the reasons that I fell in a commie-rapid hole is because I started actually looking at the evidence, supposed evidence rather, that Americans and NGOs and think tanks were pushing about their anti-China agenda. As someone who prides myself on being science-based and evidence-based, and when I was a liberal, someone who cared a lot about being correct in terms of scientific thinking, I found these “studies” terrible. Often refuting the claims that are in the title of the study itself or even if the title didn’t make that claim it was the news agency that did. Basically everything that these liberals are saying that the alt-right do to push their lies is exactly what the liberals do and they just eat it the fuck up because CNN/BBC said it instead of Fox News.
Very frustrating indeed.
As I mentioned in my other reply, that was actually what signaled me to come here and make this post because it was such a darn good video! And then they had to just go and ruin it by being like “nope sometimes we put our heads in the ground too”, and like… Ugh… damn it this is why I’m not a liberal anymore.
That’s actually literally what reminded me of it. And then my head went down this entire spiraling track and I was like, I gotta write this down.
This little interaction where you corrected the translation is why I love this platform.
It’s a simple translation misunderstanding, and when are corrected, we take it in stride and just say “oh that was silly”. Turns out it’s no big deal, and we laugh it off.
My brain is still too primed from the rest of the internet where, hypothetically, calling this out, even when it’s true, would become a big controversy.
These people have no fucking shame. It’s like they explained what American social media is to a T and then they were like, “but China actually”.
Good distinction. China hasn’t been behind in AI in general, just this new LLM stuff. But China also uses it’s AI for more useful things instead of just advertisements and recommendation engines, though I’m sure they use it there too. But yeah, China is catching up on LLMs, and fast. The chip war has affected their ability to get faster chips, and catch up even faster, but the tech they have is sufficient, and improving faster than westerners predicted (As it always does, the west is WAY to confident in itself and WAY underestimates China’s abilities)
That’s an important distinction yes, it uses a lot of smaller models added up. I haven’t been able to test it yet as I’m working with downstream tools and the raw stuff just isn’t something I’ve set up (Plus, I have like 90 gigs of ram, not… well) I read in one place you need 500 gb+ of ram to run it, so I think all 600+ billion params need to be in memory at once, and you need to use a quantized model, to get it to fit in even that space, which kinda sucks. However, that’s how it is for Mistral’s mixture of experts models too. So no difference there. MoE’s are pretty promising.
This isn’t a super surprising result. Even American companies have been talking about how China is quickly catching up in the AI space. And if Americans are admitting it, you know it’s true. Also, anybody who’s been watching the open source scene has understood that the Chinese models are very competitive. There are many many leaderboards comparing things, but Qwen, built by Alibaba cloud, is constantly at the top of the list. In fact, in one list that I’m watching, the Qwen-based models encompass the top 20.
Then, of course, they have their own closed source language models, so a little harder to test against, but by most accounts, they are right behind ChatGPT and Claude.
DeepSeek V3 is an exceptionally large model, so it’s a little hard to do direct comparisons exactly, but it’s blowing the things out of the water, and that’s pretty crazy.
I see that was his real intention all along, huh?
Since I’m not an ML engineer specifically, this article from huggingface (The worlds most popular source for all AI model hosting, and all AI data for training, think of it as github but for AI, if you are familiar with github) will do it justice more than I can: https://huggingface.co/blog/mlabonne/abliteration
Long story short, there is a small (by comparison to the total size) part of the language model that’s in charge of “refusal” if it detects you are asking something it shouldn’t answer, and you can almost eliminate that layer completely by itself. Once that is done, the model won’t refuse to answer anything, though it might still give context like “This is really illegal, but sure, here’s… (whatever you want)”. Sometimes Abliteration can take out the intelligence of a system, so you have to train it back up again.
I read The East is Still Red by Carlos Martinez, and Socialism with Chinese Characteristics by Roland Boer.
I also have in my possession Soviet democracy by Pat Sloan, but I have not read it yet, but I have heard that is a really good source for understanding the Soviet system too, of which the Chinese borrow their system from. Looking at the table of contents, it looks like it has a lot of good information. Topics like a equality of opportunity, the rights of the wage earner, What are Soviets, and some more good chapters I looks like. After thumbing through that, just now I think I need to push that to the top of my list.
I also just kind of googled around, which is honestly hard to get good information, but I was persistent and found out the different structures of the NPC and such. I also used chat GPT, which that was a while ago, so maybe things have changed, but if you prompt it correctly, like by saying " From a communist perspective, what does democracy look like under Soviet-style leadership?" Or " What are the various technical levels of government in China, including the NPC and CPC?" You can usually get good enough information that you can confirm with a separate search.
Didn’t I put this in freechat? It was a rant. You can stop gatekeeping now.