DoD paranoia on Chinese tech transfer

I recently came across a DoD report on the threat of China stealing US innovations that I found interesting.

To summarize, the report focuses on Chinese capital investment in Silicon Valley venture deals and "AI", on top of the more routine industrial espionage that takes place. Because broad economic and technological superiority is a core strategic component of overall US military superiority, the report finds this concerning. CFIUS is mentioned as being inadequate, given that it's purview is limited to larger deals. The report finds concerning the trend of broader, sustained engagement in Silicon Valley at rates that are individually too small to draw CFIUS attention.

Their broad recommendations are to expand CFIUS along the lines of the proposed FIRRMA legislation, place explicit export controls on "AI", and limit visas for Chinese students, particularly based on the fields of study. Now I would write this off as a one-off DoD propaganda effort to get FIRRMA passed, but this isn't the first time I've heard more general murmurings on the "China and AI" front from DoD affiliates. What surprises me more is how misguided all three recommendations are, and second, how blind the report seems to the actual things that China needs to acquire to become competitive with US military industry, and how to go about preventing that.

First, the focus on increased Chinese capital influx into venture deals is misguided. Venture deals provide close to zero access to either explicit IP, or to the broader process knowledge inside the companies themselves. Focusing on VC is a distraction. Second, the Chinese tech industry is already very sophisticated on the process of venture capital and new company formation in general, so its not like these deals are teaching them anything new in that arena. And minority VC stakes provide little to no leverage over the overall business decisions that the startups themselves take, so they can't strongarm US startups into making rapacious joint-ventures.

The focus on talent is correct, but the focus on visa restrictions are backwards. What's needed are streamlined visas so that those graduate students can stay in the US after they finish! If people want to leave, there's little the US government can do. Still, visa restrictions on Chinese graduate students are very much along the lines of winning the previous war. The next war is going to be employment demand for the entire global pool of PhD graduates. If China wants AI PhDs, and Chinese students are locked out of the top US universities by visa restriction, those students will still have a sizable international contingent. And whether those students are Indian, Korean, or Japanese, China can (and probably will) hire them. The only real way to keep them out of China is to have superior migration opportunities to the US. In this political climate I see no progress on that front.

Finally, the less said about export restrictions, the better. It seems like the crypto wars remixed with the flavor of the current hype cycle (AI). It will be just as boneheaded.

The report overall left me convinced that the US has no workable plan to deal with the situation. I have my thoughts on what China has to do to catch up - namely build up a first rate research university system and further depoliticize its tech industry. I have little insight into the latter, but for the former there are interesting models to copy to build out first-tier research universities and the surrounding grant-giving institutions that sustain the ecosystem. When China tackles that problem, there is little DoD or the US government can do to stop it, except try their best to outcompete it.