Skip Navigation

SouffleHuman

@ SouffleHuman @lemmy.ml

Posts
17
Comments
134
Joined
6 mo. ago

  • World’s Largest Green Hydrogen-Chemical Integrated Project Starts Operations in China

    Jump

  • The role for green hydrogen is pretty misunderstood. As an energy carrier, it's pretty bad, with a round-trip efficiency of around 30%. What it's needed for is in use cases such as fertilizer production in agriculture, and in industrial fields requiring very high temperatures like steelmaking (yes, arc furnaces exist, but they require an upfront cost that adds up on a nationwide level). In this case, creating hydrogen and ammonia with green energy during peak hours is perfect, as that energy would normally be curtailed anyway.

  • China’s core AI industry poised to surpass 1 tln yuan this year, fueled by applications, tech advances and policies: expert - Global Times

    Jump

  • Hmm, if you’re asking about me specifically, the LLMs I have on my PC are small and vastly outclassed by models hosted online. I don’t have a specific use case for them other than personal amusement and familiarising myself with the technology, and I don’t gain much from using them either.

    As for how China specifically is developing this technology, the main positive aspect is that a majority of LLMs released by Chinese firms and research groups have the model weights open under free software licenses, so everyone can download and tweak them.

    Certainly, I do not think that Chinese tech firms have the people’s interests at heart any more than other companies, but given that a push for open source AI is explicitly part of the 14th 5 year plan, I think it’s pretty clear the CPC is aware of the exploitative potential of these technologies, and is actively working to minimise the risk.

  • China’s core AI industry poised to surpass 1 tln yuan this year, fueled by applications, tech advances and policies: expert - Global Times

    Jump

  • True, I’m not blind to the rapid growth of Chinese LLMs, I’ve even got one sitting on my SSD right now. I just think that people seem to be overly focused on LLMs when there is a much broader field that is quietly advancing the productive forces, which is sadly underreported on.

  • China’s core AI industry poised to surpass 1 tln yuan this year, fueled by applications, tech advances and policies: expert - Global Times

    Jump

  • I’m not sure what you’re trying to say here. Are you saying that AI isn’t used in Chinese factories, research, ports outside of LLMs? Or are you saying that LLMs are inherently evil and no one should be developing them at all? What ‘apologia’ am I doing exactly?

  • Robotic vacuum pioneer iRobot files for bankruptcy, Chinese firm Picea to acquire 100% ownership

    Jump

  • If it ware an internet company or a fintech company, I would be more concerned, but Picea is primarily a robotics company that already makes robot vacuum cleaners, so I think we can give them the benefit of the doubt.

  • China’s core AI industry poised to surpass 1 tln yuan this year, fueled by applications, tech advances and policies: expert - Global Times

    Jump

  • Keep in mind, AI development in China is much broader than just LLMs or diffusion models. It includes traditional deep learning for science and medicine, optimisation algorithms to streamline ports and factories, embodied intelligence for industrial and humanoid robots, etc. It’s generally a much more grounded field of technological development than the Hail Mary AGI seekers in the US.

  • Socialism AI, Good idea or nah?

    Jump

  • Oh yeah, even the Chinese LLMs are pretty lib because the most of the internet (especially the English part) is dominated by liberal content, and the model makers (the Chinese ones, I mean) optimise them for coding or math performance without caring much about ideology.

    There is definitely an opportunity for someone to fine-tune it with Marxist content, but the prerequisite for this is to build large enough datasets with the right information and analysis. This is why stuff like Prolewiki is so important.

  • Socialism AI, Good idea or nah?

    Jump

  • If done well, I think LLMs could be a pretty decent entry point for newbies. They don't need 100% perfect dialectical analysis, they just need to be pointed vaguely in the right direction. As long as you can train it to emphasize the importance of reading actual theory rather than just relying it on information, it wouldn't necessarily be a bad thing.

  • Is it possible the chuds are shitting themswlves a bit too quickly?

    Jump

  • Well, the American right is not really a unified block with common material interests, but a hodgepodge mix of QAnon conspiracy theorists, cultural warriors, neoconservatives, Christian Nationalists, corporatists, etc. It’s no surprise that as a whole they are self-sabotaging and dysfunctional. The big surprise really was that Trump was able to hold all of them together for so long, but even that looks increasingly untenable as his popularity wanes.

  • America's Panic Manual

    Jump

  • That’s true, but I’m worried about the thousands of Machado wannabes dreaming about becoming part of the comprador class. Also, it feels like the pink wave is faltering and a new tide of fascists are being elected.

    I believe the litmus test for your question is Brazil. They have the scale and power to meaningfully counteract US imperialism on the continent, but also the resources to sustain US hegemony if they were to become a glorified mining pit. I hope they can receive enough support from their BRICS partners to achieve the former.

  • America's Panic Manual

    Jump

  • The goal is to “deny non-Hemispheric competitors the ability to position forces. Competitors cannot position other threatening capabilities. They cannot own or control strategically vital assets in our Hemisphere.”

    I believe this is the linchpin of the US' strategy. Latin America contains vast amounts of resources, a large labor pool, and a strong comprador presence in almost every country. It also happens to be in an area where the US has a near insurmountable logistical advantage compared to their main competitor (China). If the US could dominate the region and extract their resources, it could feasibly maintain its hegemony for quite a while.

    Mexico is the largest producer of silver, with nearly 25% of global production. They are also in the top ten producers of gold, copper, zinc and lithium.

    Brazil is the largest producer of Niobium (around 95% of global production!), has the second-highest reserves of rare-earth minerals (after China), the second highest production of iron ore (again, after China), and the second highest hydroelectric production (also after China).

    Also, there's the Lithium Triangle, a relatively small region in Argentina, Bolivia, and Chile that contains 54-58% of all Lithium reserves in the world, and it's a vital component in things like battery production.

    There's a lot more that isn't said here, like Venezuela's gargantuan oil reserves, but the point is that the American imperialists are no doubt drooling at the idea of plundering the region and extracting unimaginable amounts of super-profits. Fortunately, the US is still unlikely to pull a significant portion of their forces away from Asia (containing China is the whole point), and they can't pull out of Europe or West Asia instantly either. But still, I do genuinely fear for Latin Americans, and I do hope they continue to resist American pressure and stay vigilant.

  • China leads world in per capita electricity generation

    Jump

  • It’s a climate thing, hydro worldwide decreased because there was just less water passing through dams this year.

  • Chinese SMIC Achieves 5 nm Production on N+3 Node Without EUV Tools

    Jump

  • Huawei actually already has an EUV prototype, and is believed to be currently in the process being tested and undergoing trial ruins. If all goes well, mass production should begin either next year or in 2027. It apparently uses a different approach to ASML's, so there is a bit of risk, but it is theoretically sound and I have confidence that they could figure it out.

  • Former Chinese senior banker Bai Tianhui executed for taking US$155 million in bribes

    Jump

  • Yup, mostly it’s a two year reprieve where the the execution would only be carried out if there are additional violations or if the convicted was uncooperative. According to the article, the court said:

    “The amount of bribes received by Bai Tianhui was extremely large, the crime’s circumstances were particularly serious and the social impact was particularly severe.”

    So they decided to go straight to capital punishment.

  • Former Chinese senior banker Bai Tianhui executed for taking US$155 million in bribes

    Jump

  • Huh, no two year reprieve for this case. Must be particularly severe.

  • China's Heavy-Duty Electric Truck Fleets: A New Era For Mine Operations & A Possibility For Transporting Cargo With Zero Fuel Costs?

    Jump

  • Yup, it’s so cool how we can now make use of the gravitational potential energy of the cargo to charge up our vehicles now. Of course, the key requirement is that the truck needs to carry stuff downhill (and stay empty uphill) for this to work, but even so, electric vehicles are just much more efficient at converting energy into propulsion than ICE vehicles in almost any situation.

  • In today's geopolitical circus, Japan fumes as the Trump administration remains conspicuously silent on its escalating dispute with China

    Jump

  • Announcing: Thaura - Your Ethical ChatGPT Alternative

    Jump

  • Most AI platforms use massive models with trillions of parameters that activate all their computational power for every single query.

    The first part is probably right, frontier models are likely around a trillion parameters total, though we don’t know for sure, but I don’t think the second part is correct. It’s almost certain the big proprietary models, like all the recent large open models, use a Mixture-of-Experts system like Thuara, because it’s just cheaper to train and run.

    While traditional web research might take you 10 minutes of clicking through pages and consuming cookies from major search engines, Thaura uses a fraction of the energy and provides the same information.

    This part is pretty misleading. It is very unclear how much an LLM query compares to a search in terms of energy use, but even assuming they’re about equal (most estimates put LLM queries higher), the LLM also has to do their own web searches to find the information, if you’re using it for research purposes, so that point is fairly moot. Also the “consuming cookies” part isn’t really an energy problem, but a privacy problem, so I’m not sure why it’s used in this context.

    Thaura uses a “mixture of expert” models with 100 billion total parameters, but only activates 12 billion per query.

    Going to the actual website, it does credit the “GLM-4.5 Air architecture”, but the article doesn’t mention GLM, or the company behind it (Z.ai) at all. Given that this is likely a finetune of the GLM model that was freely released, it feels weird how the Thaura team seem reluctant to give credit to the Z.ai team.

    These companies are often controlled by US-based corporations whose political stance supports occupation, apartheid, and Western hegemony - not human rights or global justice.

    Reading below and also looking at their website, the hosting and inference is done by US firms (DigitalOcean, TogetherAI) in datacenters hosted in the EU. That’s not inherently bad from a privacy standpoint due to encryption, but it does feel disjointed that they are railing against US firms and western hegemony while simultaneously using their services for Thaura.

    While I don’t think the Thaura team had bad intentions in fine-tuning their model and building the service, I feel that this is a pretty misleading article that also doesn’t give any significant details on Thaura, like it’s performance. They also haven’t given back to the community by releasing their model weights, despite building on an open model themselves. Personally, I think it’s better to stick to Z.ai, Qwen, Deepseek, etc, who actually release their models to the community and pretrain their models themselves.

  • We should improve society somewhat

    Jump

  • Meanwhile the Centrist: Wow, both of you are equally awful!

  • Robots and AI Are Already Remaking the Chinese Economy

    Jump

  • Thanks for the clarification. I do agree that China's development in AI / deep learning is significantly broader than the hype in the US, which translates to a more robust development in the productive forces. The other thing I wanted to ask is the bit about China's shrinking population. Given the indispensable value of labour in the economy, is that a legitimate problem in the long run? Or is just the latest evolution of the western "China is Collapsing!!!" narrative?