Interesting if true - that Moonshot can train a ~3T SOTA model on only 20K NVIDIA GPUs, while others like Musk (who freely admits to distilling OpenAI's models) struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs).<p>I guess at least partly a reflection of all the optimizations in the Kimi 3 architecture.<p>In the recent leaked DeepSeek investor meeting, they also mentioned only having a 20K GPU cluster (unclear if NVIDIA, or Huawei).
It kind of confirms a hypothesis I have that the next phase of AI development will be about getting <i>smaller</i> (in terms of model size and compute), because smaller is more capital efficient for training (allowing faster iteration and more iteration cycles for a given amount of capital), allows for denser inference (more inference for a given amount of compute hardware), and allows for more edge inference applications.<p>The goal will be to develop smaller models with more efficient architectures, that have similar or even better performance than larger models.
Yes, continual learning is currently a hot topic since this is what will allow an "AI intern or new employee" to learn on the job rather than be stuck on groundhog day, or pre-trained for every eventuality - needing to anticipate all the quirks and proprietary knowledge of every customer!<p>Continual learning tends to imply individualized models, else there is no data privacy (the secrets learned on the job at your company now being available to your competition), which really turns the current AI business model of a single centralized model served to everyone on it's head. If every customer has a different model that essentially means the end of batch processing with the same weights loaded into the GPU.<p>The direction this suggests is a move away from centrally served common models to locally served individual ones, which generally requires them to be smaller, even if some larger companies may be willing to invest in beefier hardware.<p>I think this is at least in part why the AI companies are trying NOT to implement true continual learning and see if they can instead finesse it continual compacted(?) memorization instead, since then it's "just" additional context that needs to be recalled and fed into every request, not weights that need feeding into the GPU. I don't think memorization is any substitute for learning, especially learning of practiced skills, but since it's far easier to implement, and non-disruptive to the cloud-based API business model, this is what we will see first.<p>The recent news of NVIDIA' investment in Sutskever's SSI has a tiny hint of this also, talking about SSI advising on NVIDIA's future architectural direction - apparently pushing it in a different direction than current (cloud-based, pre-trained) models. NVIDIA may be quite happy to see a move towards local models.
This is likely what Google has been doing, as it suites them best to have small fast models rather than hulking slow giants. Lower intelligence but way more ability to serve. Anthropic would probably need a datacenter the size of a small country to serve Fable on a Google search/services scale.
Getting smaller is how the last revolution in computing happened.<p>A VAX 11/780 was good, but an 80386 was a lot better, since the latter could run on 3 AA batteries and the former needed 6,000 watts of 3 phase.
It's not that surprising to me. Most of the innovation in Chinese models has been in efficiency gains and optimizations. K3 coming from the factory in MXFP4 weights is a pretty relevant factor. Big performance gap probably also due to Moonshot doing QAT. Throw in the fact that Musk has easier access to compute, and I think you have your answer on the disparity.
One does wonder whether there has been an expiration of the actual weights of Opus at one point.
> <i>Musk (who freely admits to distilling OpenAI's models) struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs)</i><p>Are you sure they are using all of their compute on training? Didn't they rent out a ton to other AI companies?
Genuine question. Is there a time factor part of that equation?
> Musk (who freely admits to distilling OpenAI's models)<p>Source?
My thesis is still that model building has no moat. Folks continue to migrate around between the big labs. There is a lot of value in having good taste around the harness and how the models are used. The medium to long term winners will be the folks that control the compute.
You can only control the compute if it is not commoditized, which at the current moment, it is not, leading to industry players having 75-80% margins.<p>Someone will find a way to make cheaper compute, and since nothing fundamental changed there (LLM didn't change how silicon were made), that's bound to happen.
<a href="https://archive.is/vgmMB" rel="nofollow">https://archive.is/vgmMB</a>
I read this as if they purchased it from Alibaba the marketplace.....
K3 is natively trained to mxfp4, if they cannot get a hold of Blackwell chip, it is meaningless. Hopper does not do native 4-bit floating math.<p>Either they have Blackwell with native 4-bit floating math, or they use have Chinese domestic NPU that support mxfp4 natively.<p>The article’s statement does not make sense.
> Either they have Blackwell<p>From the article:<p>A person familiar with Moonshot’s procurement strategy confirmed that the company does indeed have a channel for accessing Blackwell processors via Southeast Asia. They didn’t specify whether this was a rental channel...
Huawei A910 is supposed to have native support for MXFP4. The problem is that huawei doesn't have the capacity to build more GPUs right now. Those yield numbers must be really terrible.<p>Personally, I think 20K nvidias is a stop gap solution because they really don't have the capacity to serve their models to earn any money right now.
What do you mean by "it's meaningless" can you expand?
I would not minimize the extent to which AI progress depends on things like the effort that goes into training material acquisition and data labeling. In those areas, improvement is a direct function of how much you invest.<p>I wonder how much spending is motivated by the phenomenon of sudden emergent performance in LLMs. Clearly some people who are smarter than me expect something like emergent AGI, or at least they think the odds justify spending whatever it takes to see if that would happen.<p>That leaves a lot of room for efficient aggressive followers.
<a href="https://archive.md/vgmMB" rel="nofollow">https://archive.md/vgmMB</a>
It's also built on the back of downloading all the content of the top ten sites on the list, the fact that all the independent content sites with their "pre-war" steel haven't been acquired at top dollar by U.S. companies shows the existential threat / stop China posturing is a total joke.<p><a href="https://archive.is/AvWWX" rel="nofollow">https://archive.is/AvWWX</a>