jason Hsu

@JensenHuang @nvidia The United States has never dominated in technology by locking its platforms away. It dominated by making its platforms the ones the world’s developers chose to build on. So the answer is to #outbuild China. Open-weight is the way to go! nytimes.com/2026/07/23/opi…

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…

For those gorging on free or cheaper Chinese AI models, be careful. Don’t you understand that Chinese companies will raise their prices or restrict their models? Don’t you understand that it will be very complicated and expensive to replace your Chinese AI model when this happens? Don’t you think this is by design? Don’t you understand Beijing did took similar steps with solar panels, telecom and equipment, and critical minerals? Great op-ed by @augama

@isaacstonefish Thanks for sharing @isaacstonefish let’s catch up sometime

This does not threaten OpenAI or Anthropic. Their business is the frontier: the hardest, highest-value work, sold to customers who will pay for the best. The only question is whose model sits underneath. It should be an #American one

Great to publish Op-ed @nytopinion on #China’s #AI strategies, #openweight models and how U.S. should respond. nytimes.com/2026/07/23/opi…

Fund a contract to train an advanced open-weight model, with the weights permanently open. Finance the #full-stack abroad with #trusted allies — chips, storage, compute — not the model alone.

The United States has never dominated in technology by locking its platforms away. It dominated by making its platforms the ones the world’s developers chose to build on. So the answer is to #outbuild, not to block:

dependence is the strategy.China ran the same sequence with solar, telecom equipment,critical minerals: build the world’s dependence on something cheap, then restrict access once the dependence is real. A country that cannot switch off a foreign model has handed Beijing a grip.

Alibaba’s #Qwen became the most downloaded model family on earth, and why developers have built more than 100,000 specialized models. And much of that adoption is American. Airbnb’s customer service agent runs on Qwen. Cursor built Composer 2 on a base licensed from #Moonshot.

Here is what I think gets missed in this debate. #Openweight models are not a category of software. They are a #foundation. When a company publishes a model’s weights, anyone can download it, run it on their own hardware, inspect it, and retrain it for a specialized task.

My argument is: China spent three years giving its most capable AI away for free, and is now moving to restrict access to its most advanced models. Washington’s instinct will be to answer with a ban. That would be a mistake.

China is beginning to restrict overseas access to some of its most advanced AI models. The US should not respond with a ban of its own. America and its allies should build and deploy their own open-weight models, @augama writes in @nytopinion. Read: nytimes.com/2026/07/23/opi…

Roughly 95 percent of the world's data, from financial transactions to daily messages, travels through undersea fiber-optic cables. They are thin, fragile and almost entirely unprotected from either seismic activity or deliberate sabotage by bad actors. China has sabotaged cables near Taiwan, and Russia has damaged cables in the Black Sea. Now, cables under the Strait of Hormuz are at risk. The US needs punish foreign persons and vessels responsible for sabotaging undersea cables and encourage allies to harden their infrastructure and speed up repairs, Jason Hsu writes. Read @augama's latest in @thehill: thehill.com/opinion/congre…

Taiwan’s drone strategy should ultimately be about more than drones. It should be about building a new model of asymmetric deterrence—one based on scale, speed, resilience, and continuous innovation.

I was glad to speak with @TaiwanPlus for the final episode of its three-part series examining how the wars in #Ukraine and the #MiddleEast are reshaping modern warfare—and what those lessons mean for #Taiwan. youtube.com/watch?v=EWdGNP…

The challenge is to translate those capabilities into a resilient defense technology ecosystem that connects the military, private sector, and frontline operational needs.

Recent conflicts have shown that technological advantages can disappear quickly as adversaries develop countermeasures. The side that can learn, iterate, manufacture, and deploy faster will have a decisive advantage. youtube.com/watch?v=EWdGNP…

The key is not only to build sophisticated platforms, but to create systems that are affordable, scalable, rapidly replaceable, and constantly adaptable.

For Taiwan, developing a strong indigenous drone ecosystem is therefore not just an industrial policy objective. It is a national security imperative. youtube.com/watch?v=iWBQs2…

One of the clearest lessons is that #drones are no longer simply supporting assets. They are becoming an essential part of how militaries conduct surveillance, targeting, communications, logistics, and strike operations.