Former Meta PM Xiaoyin Qu predicts US and European enterprises will adopt Chinese AI models to self-host on private GPUs

"American and European enterprises will ditch OpenAI and anthropic and adopt Chinese models." I love fireworks, Nebius et al But they don't have the scale or compute required to support this

"They can host Chinese models under their own GPUs" => next blogpost will be how this isn’t as straightforward as that.

@Dorialexander Been thinking about this a lot, looking forward to reading your thoughts on it.

@rohanpaul_ai That’s not going to happen. The U.S. government won’t allow it. They’ll ban any Chinese model that’s on par with Mythos/GPT-5.6.

@JuanEcheverrria @rohanpaul_ai how？the chinese models are free and open sourced

@Dorialexander Don't get me wrong, it's doable. It's just a lot harder than it sounds and exponentially harder when you are serving it for more than one person

@Dorialexander Can attest , it's not easy

@_xjdr If even you say it…

@rohanpaul_ai And shes right

@xlr8harder Incidentally also matters to run this on HPC for synth…

@_xjdr No but that's really the key thing: average large co might have the compute, just not the skills.

@Dorialexander HPC solution for this is much harder. Slurm is not shaped well for this workflow. I might have a useful tool for that environment, will DM you.

@ThomasYiwen @JuanEcheverrria @rohanpaul_ai They almost banned deepseek like a year ago, making it a crime in America to use it. I think the real motivation is to keep the AI models more distinct from one another, rather than having them be effectively siblings due to cross-model distillation.

@ThomasYiwen @JuanEcheverrria @rohanpaul_ai The reason to do that would be to mitigate/limit the damage of any serious failure condition. Model training is durable. Serious flaws are often discovered late, and you can't easily undo mistakes.

@usmanconsulting @rohanpaul_ai You can’t even put a website together, Usman, please don’t try and lecture others on AI. Ain’t nobody taking you seriously

@rohanpaul_ai Man you're posting great news today

I hope all closed source gen ai go bankrupt

they train with public data, books and protected data (articles/blogs) then has the nerve to add a paywall

@rohanpaul_ai Very interesting. Curious how all this will turn out 🫡

@JuanEcheverrria that's a possibility as well

@rohanpaul_ai Yeah, the US should be more careful with Communist Troyan horses.

@zephyr_z9 企业迁移大模型不是换个 API 这么简单，合规、算力、服务稳定性都要兜底。中国模型有性价比优势，但承接全球企业还差基础设施。