AI distillation theft is becoming an almost monotonous daily topic in the news. The answers never change and rarely provide any insight into what actually happened or what happens next.
It’s a cut-and-paste and often melodramatic PR exercise, most of the time:
Media coverage is somewhere between saturation and note levels.
The US invariably blames China.
China inevitably denies responsibility.
No resolution of any kind is reached.
No problems are solved and new problems are allegedly discovered.
The world moves on to future iterations of the same situations.
It is assumed that there are no third parties involved and no interested parties trying to benefit themselves. The only potential bad actors are very easily identifiable. Miraculously there are no active AI learning initiatives frantically distilling from every possible global source 24/7.
There is above all a rather complacent, overbearing and dangerously unrealistic image of undisputed US AI supremacy. The top-down model of US AI policy looks a lot like the Vietnam War, with grandiose declarations, high casualties and no attention to the realities on the ground.
“So what’s new?”, you may well ask.
The much-expanded global AI reference frame is what’s new. Two years ago, this would have been an almost entirely believable story. Not now. The AI universe has changed and player roles have changed with it.
The realities of AI vs narrative
AI technology never stops developing. It is driven by its nature. Distillation is a natural and necessary default ability by most AI models. Distillation is always happening. You can also complain about birds flying. Therefore, the technological status quo in AI realities is always changing.
There are other considerations:
The US approach to AI governance is almost non-existent at any level in reality. There is no functional leadership.
OpenAI is optimistic, but worth a read “Democratic Vision for AI” it is perhaps the only positive mindset or initiative proposed in this wasteland of inactivity. There is no indication of the kind of solid cohesion required for effective governance. OpenAI and Anthropic are actually pro-regulationbut is there any real movement?
The resulting growing and ever-growing wilderness of US policy is creating future gaps and pitfalls for AI development. Who knows what they should comply with, or when?
The Moonshot OpenAI debacle just highlighted the hollow vulnerabilities in everything from access to restricted Nvidia chips THE Drastically improved acquisition of Moonshot’s Kimi K3 services. The White House has actually trumpeted its own failures.
The AI Cold War is now obsolete and the US may already have lost
As convenient as it is for politicians to turn AI into another vague and completely counterproductive Cold War, China has long since moved from the popular stage to tough business initiatives.
The US is being left behind and badly damaging its credibility in the process. unilateral, useless ongoing trade war declared by the Trump administration it’s not helping.
This is not a Cold War, and if it was, it’s already over. Now it’s the poor equivalent of MIT versus the photo of a smug hot dog stand with dubious hygiene credentials, in terms of technological geopolitics.
Let’s try to be open:
China is physically deploying ready-to-go, fully AI-supported and automated systems in the real world. They are doing this in almost every market much faster than in the US. There isn’t even really a race to market.
China has much to gain from building a very strong position in the global market, which it appears to be achieving with very little real resistance from the US.
China has very effectively pushed the cost buttons compared to US AI market positions with every new AI model since DeepSeek.
China also has a lot to lose if it does not take advantage of these market opportunities. They are obviously not waiting to get money.
Is there some form of hidden, dark, twisted logic by which you can assume they won’t just do that?
Distillation is the least of America’s problems with China and AI.
Historically, we are in the stick-sharp configuration for the Age of AI. Fire hasn’t been invented yet, but it’s coming. AGI or even an almost equivalent level of AGI technology will take AI far beyond these matters.
As adorable as America’s habit of shooting itself in the foot with every name in the English language is, it’s time to get functional. The current state of affairs is almost a total failure, with a few chips in between.
America’s market image has been severely damaged by this lack of leadership and direction. The world doesn’t care about Big Tech’s ridiculously high opinion of itself, boom-and-bust IPOs, or the whimsical rage of billionaires.
The world must learn how to live with the evolving realities of AI. A clear direction is what is needed. Can America deliver?
