America Is Building the Great Wall of AI — And Trapping Ourselves Inside

For decades, America criticized China for walling off its digital economy. However, when it comes to AI, the roles may be reversing. In this week’s Shimmy Says, Alan Shimel argues America is building its own Great Wall of AI — and the wall may end up trapping American developers, researchers and businesses on the inside.

The Argument

Chinese labs are releasing capable open-weight models like Kimi, Qwen and DeepSeek. Moreover, they compete on price, accessibility and developer control — not just benchmarks. American AI incumbents see a threat and are asking Washington for protection.

Alan says take the security concerns seriously. Therefore, reject the proposed solution. The real risk is not Chinese AI getting in. It is American developers getting locked inside a protected market while the rest of the world builds with whatever models work best. That is the Great Wall of AI.

The Indispensability Trap

Alan connects the argument to his forthcoming book, The Indispensability Trap. In addition, he explains why the AI incumbents want the rewards of becoming indispensable — without accepting the market and regulatory pressures that always follow. Open models, government policy and public demand for access are not creating that trap. They are springing one the industry set for itself.

Chapters

00:00 From China’s wall to America’s wall
02:00 Kimi, Qwen, DeepSeek — the open-weight surge
04:30 The Indispensability Trap explained
07:00 “Full AI communism” — what OpenAI’s Dean Ball actually said
09:00 Commoditization vs. communism
11:30 Anthropic’s fraud allegations against DeepSeek and Moonshot
13:30 Punish theft, not competition
15:00 Why the Great Wall of AI can’t work
17:00 The asymmetric cost of walling America in
19:00 The likely future — frontier plus open-weight stack
20:30 Protect IP without protecting artificial scarcity