I think it's as simple as looking at the incentive structure. The Chinese gov't has incentive to kneecap US monopoly on frontier models. It makes sense for them to continue down this course if it strengthens their position.

Chinese gov is not even needed. Deepseek was already profitable before this whole thing started. They even had the GPUs.

how profitable though

The Chinese govt has a strong incentive to break dependence on the US for AI needs and build domestic models, but not to spend trillions to subsidize these models for the rest of the world.

Agree but I think they'll spend whatever they can to subsidize and break dependence on US models. Whatever isn't spend on US models will be spent on chinese models. Europe will remain a small single digit % unless they get their act together.

The Chinese gov't and also the rest of humanity...

I'm not sure how everyone in the US forgot that monopolies are bad

Closed models doesn't mean monopoly. Nothing stops new competitors, in fact there is strong competition right now.

Who's got a monopoly though lol?

There is plenty of room for open source models that "require" a subscription to be used or obtain working updated binaries. Same as any open source platform.

I'd be happy to pay a small monthly fee to license the model to run locally. I'm already paying for Claude, GPT, Gemini,etc.

Right, this is the classic VC playbook and people should really know better. They use capital to undercut competitors on cost to gain market share, then slowly crank up the costs until they are (hopefully) profitable. It makes 0 sense long-term to spend tens of millions on training a frontier model and releasing it for free for someone else to host on their GPUs. It is naive to believe that when the capital starts to dry up and the VCs are looking for a profit that the models will continue to remain open-weight.

Chinese labs are a bit different since they are somewhat state-funded, so I'd expect them to shift to a model where Chinese models are hosted on Chinese infra.

> I'd expect them to shift to a model where Chinese models are hosted on Chinese infra.

You assume these Chinese companies need to make a profit. Nothing is really private in China, nothing that truly grants power, anyway.

CCP can just write off these loses as defense budget, which they basically are.

I bet that, all things considered, training 10 new Chinese frontier models costs _significantly_ less than a single modern fighter jet.

Which on itself is wild. The main issue with the discussion isn't a rift in capabilities, but a rift in the valuation model itself.

If China can match the pace and produce frontier models at a fraction of the cost of US alternatives, even without access to SOTA hardware, what is the competitive advantage for the current valuations of Anthropic and OpenAI?

The Chinese ecosystem is largely built on stealing intellectual property from and/or distilling of US closed source models.

I’m not getting into the ethics/comedy aspect of it, but let’s not pretend this isn’t the case. Plenty of evidence, the motive is obvious, and the numbers are in plain sight (valuations, salaries, etc).

As long as this continues I think closed source will continue being a few steps ahead, but the steps will probably get smaller over time.

Having said that, this is already priced in, the market predicts this gap will be large enough for the US companies to profit from (astronomically).

Basically. The article doesn't say in what way they're winning or what the US should do.