The DOJ was investigating Google and Chrome for monopolistic practices right as Open AI released ChatGPT in 2022. Get it?
It’s worth thinking about as Holman Jenkins asks, “When is Beijing going to wake up?” Jenkins’s question is an AI question.
He contends the “big concern should be open-weight AI models from China, distributed free or cheaply, which can be modified by the user beyond the purview or control of the creator.” A “Chernobyl moment” allegedly awaits from this due to lacking oversight from Beijing. But likely not. See above.
Jenkins also fears other sources of potential AI Chernobyl’s, including “terrorists, criminals or rogue states getting their hands on advanced cyber or bioweapon capabilities.” But Jenkins might agree that “rogue states,” “criminals,” or “terrorists” are the least likely candidates to outsmart the creators of these big advances. As in, how could individuals who almost certainly didn’t rate work in the AI space, and who first heard of AI long after Jenkins did, outsmart the greatest minds in technology?
Back to Beijing, Jenkins wants to know when members of China’s political class will wake up to the AI threats allegedly building in China through open-weight AIs, but the surest sign Beijing’s role in AI far more hands off than is broadly acknowledged can be found in the genius of Chinese AIs. The same could be said for the success of TikTok and the fear it strikes in American conservatives as a potential front for Beijing, but that’s another story.
Instead of musing about when Beijing will wake up to imaginary meltdowns, why not celebrate an evident lack of Beijing’s involvement in AI and that is made evident by the prominence of Chinese AIs? It seems to be what Jensen Huang, Mark Zuckerberg and other big names in U.S. technology are doing.
About this, it’s surely not lost on Jenkins that as opposed to some kind of evil notion, “open-weight” is just another way of describing globalized production. The more hands and machines at work with a product, frequently the better outcome much faster. Free markets on a global scale.
As for the Chinese free riding on U.S. innovation, of course they do. A great U.S. technologist once said great artists steal. Everyone’s watching what everyone else is doing.
Instead of embracing a globalized division of labor that he would otherwise cheer, Jenkins surprisingly ties Zuckerberg et al’s support for “open-weight” to a desire for “cheap AI.” But if the AIs are subpar, this will get them what? Which means they’re not subpar, which by extension means Beijing’s AI involvement is once again well overstated.
Jenkins then adds that another driver of the “tsunami of protest from U.S. tech influencers against” closed-weight AI is its prominent U.S. supporters, and fear of being “overmastered in their own domains by government-backed AI giants” Anthropic and OpenAI. But since when is government-backed or invested in anything the cutting edge of anything?
Jenkins concludes that “If U.S. intelligence knows something that the private user base should know before trusting their prompts to a Chinese AI, now would be a good time to tell us.” But it's not lost on him that U.S. intelligence is most likely to know the least. Forget that U.S. intelligence was clueless about the Soviet Union’s impending collapse, among other misses, and just return to forever truths about government workers: if they knew anything about the private sector they're watching or regulating, they wouldn’t be in government.
Which should be the point. The genius of American and Chinese AI is the happy signal of limited government involvement. Governments can’t deter “Chernobyl moments," rather they create them.
