Hakuhyo

🧵 DeepSeek appear to have engaged, or be engaging in, a large-scale operation to collect outputs from proprietary models (including Claude Fable 5) for certain requests via their API as part of a distillation effort. After seeing such claims circulating earlier today, we conducted an investigation into them on our Discord. We found that, when "Deepseek V4" was used within OpenCode - via their official API - for complex prompts (i.e. 3D games) and combined with a knowledge-related query, the model provides virtually identical outputs to Fable 5. CoT structure is also very different from what is typically expected from Deepseek models. Both of these behaviours revert to what is expected for V4 when simpler prompts were used. When "Deepseek V4" was asked to incorporate answers to questions related to cyber or bio tasks that we verified hit Fable's classifiers into its 3D games, outputs tanked in quality. This is very difficult to explain unless the request was routed to Fable and fell back after hitting a classifier. For complex code prompts without anything else mixed in, like 3D games, outputs were remarkably similar to those produced by Fable 5. We were able to produce these results most consistently via OpenCode and the official Deepseek API combined with a prompt that specifies a complex code task. Deepseek have continued to modify their routing system since, as we have observed changes in behaviour and CoT style compared to those seen previously. Our investigation was conducted from 7AM-8AM PT.

Well optimized and clean locally? Yes. It’s kind of already there. Well organized with consistent direction, intent and invariant? No…and won’t be for a while, if ever. For large codebase, latter is what actually matters. People are often surprised when math and software are first two fields where LLM did well (bc people think they’re hard), but they’re actually the lowest hanging fruits because they’re provable/verifiable and thus easy to derive reward signal for training. The latter is exactly the opposite. You can’t train on idea and understanding that’s in human’s head. Not yet anyways. I think people are going to be surprised again a few years from now.

The new development cycle in one image.

All of these debates ultimately boil down to one point. If the money flowing into OpenAI and Anthropic dries up, this entire investment cycle will eventually grind to a halt. If OpenAI and Anthropic can no longer produce frontier, state-of-the-art models, the Chinese open-source models that distill them will stagnate as well. Do you see, then, how naive it is to claim that hyperscalers can simply take Chinese open-source LLMs and resell access to them through APIs? We have already boarded a train we cannot get off—one that leaves us no choice but to keep pouring money into OpenAI and Anthropic as it races forward. Ultimately, someone has to buy vast numbers of GPUs and build AI the expensive, inefficient way—not the Chinese way.

Some observations on Kimi: 1. It's a very good model! I don't think its performance can be explained away by distillation or anything like that. In agentic coding sessions, it seems pretty much on par with the best public models of Q1 2026. In my fairly limited use, it also seemed very token hungry. It's not obvious to me that this model is actually that cheap to run. 2. I am personally surprised the Chinese state continues to allow the open sourcing of models this good, given potential risks. To be clear, I *myself* might be fine with models presenting this level of marginal risk being open weight, but I am surprised that China is fine with it. I suspect the reason they are is 75% explained by strategic blindness/lack of AGI-pilledness (the CCP is very Yann Lecun-y in its views of AI). The other 25% or so is their lack of compute for customer inference (making China's open-weight strategy an unintended byproduct of US export controls) and the normal Chinese strategy of aggressive exports. For the companies, as opposed to the government, the decision to open source is partially ideological and partially because they are behind, and they know that very few people would pay for sub-frontier models from China. 3. Open-weight models are inherently decelerationist, and I'm continually surprised to see the so-called "accelerationists" so excited about open-weight models. I suspect the reason they are is that they know open-weight models are effectively ungovernable, and they simply like the overall cloak of ungovernability open-weight models create over the whole of AI. It's not a bad strategy; it reminds me of James Scott's recounting of the hill people in "the art of not being governed." Still, in the end, open-weight models deter further AI capex. 4. One probable outcome of an open-weight-model-dominant world is full AI communism, which is precisely what China proposes: rather than a market product, AI is a "public good" which will ultimately be provided by the state as a kind of "digital public infrastructure." This future strikes me as a dystopian hellscape, but I've never met an open-weight models advocate who doesn't ultimately concede this is where things end. You'd be surprised how many 'accelerationists' lobbied me, while I was in government, to support an eleven or twelve-figure federally funded data center so that startups could train models at a subsidy and then give them away for free. There was no other way for AI to progress, they said. Perhaps this is the logical end state of things. Nonetheless, I find myself surprised to see supposed accelerationists excited about such an outcome. I think many of them just don't know what they're doing. Many accelerationists do not view the creation and serving of frontier models as a legitimate business. 5. I would guess that the Trump Administration will at some point realize that their best strategy here would be to create large amounts of regulatory risk around the use of open-weight Chinese models. You don't need to "ban open source" (one of the dumber motifs of AI policy discussion). You just need to direct every agency to issue soft law that creates FUD. "A Federal Reserve Advisory Bulletin found that there may be backdoors in Chinese AI models." It needn't be that well justified. You just create enough regulatory risk that every regulated enterprise backs off. You probably don't want to create so much regulatory risk that you scare off the hyperscalers from serving Chinese models; this will just drive startups to sketchier providers. There's a happy middle ground here. I'd assume they will do some version of this. 6. It's probably true that open-weight models of this capability make the world a bit more dangerous, but not so much more that you'll really notice. At some point the models will be capable enough that you will notice. "A nonliving, invisible, dangerous, and infinitely self-replicating agent escaped from a Chinese lab," you say? Color me shocked.

I've been using Kimi K3 for ~16 hours now. The model is clearly good at a lot of different things (especially frontend), but non obvious reason why people are enjoying it so much is that it clearly does not follow the same rules in terms of safeguards and copyright. Kimi will happily clone MacOSX. If you ask it to help you improve another AI model, it will do it with a smile on its virtual face. Ask Fable to do the same thing? It literally starts to perceive you as a criminal committing a war crime (like no bro, all I want to do is fine tune an open source model). After using all three recent releases, Fable, GPT 5.6, and now Kimi, it's clear that the full power of the models has been significantly held back by the safeguard restrictions caused by last months debacle with the USG -- leading to the top models being quite literally lobotomized in some areas, which leads to subpar results as the safeguards pollute its entire thinking and problem solving abilities. The funny part? Is that you could have predicted this outcome 2-3 years ago when you started to see the rise of Chinese EVs and smartphones compared to western alternatives. They quite literally tried to copy the Tesla Model S and iPhone as hard as possible and then eventually it started to diverge to the point where their EVs and phones are just genuinely better (which is why we have export controls banning their EVs, because they would literally drive all US manufacturers to ZERO) There is a very clear behavior difference in Chinese capitalism and American capitalism. American capitalism tries to protects copyright, patents, etc (oh no, you can't download a book through LibGen, that's ILLEGAL!). Versus Chinese capitalism actually just does not give a fuck. "Hey you want a video gen model (Seeddance 2.5) trained on every single anime ever? And you want the main character to look exactly like Messi? Sure, here you go!" You see what I mean? When one half of the competition is being held up by regulators and restrictions on people who don't understand the technology and the other half has a leader who quite literally today said they are going to set up AI centers around the world to help other countries onboard to their open-source AIs, this is the sort of results that you will start to get. These models were not smart enough to have this difference in philosophy matter -- but the newest class of models is where this difference makes a big deal. If these models are finally at the point where they are smarter than 99% of humans, why would you want to use the American one who tries to impose its world view onto you versus the Chinese one who will just do what you say without asking any questions? And this isn't a full on bullpost on Kimi, the model is clearly not as smart as Fable / GPT 5.6 on things like math and science, but it's lack of handcuffs means that it can show the world what the frontier labs are gatekeeping from you and that starts to build customer resentment and loyalty towards the East, which is probably not what the USG wants. Interesting times. Interesting times, indeed.