Two stories dropped this week that look like opposites but are actually two faces of the same anxiety. Current AI, a nonprofit, is racing to build open AI infrastructure that leaves no culture behind. Meanwhile, The Atlantic reports that Beijing is doubling down on a deliberate strategy to keep out Western chips and tools entirely. Both moves are, at root, about who controls the stack beneath intelligence itself.
The Open-Access AI Bet and Its Historical Parallel
Current AI's pitch echoes the original promise of the World Wide Web: a commons that no single government or corporation owns. It is a direct counter-thesis to the consolidation happening at OpenAI, Anthropic, and Google DeepMind. The nonprofit framing is not accidental. Brewster Kahle has been making a nearly identical argument about libraries and public AI for years. As Kahle told Culture Slop, the global brain only works if it is genuinely public: closed infrastructure is just another way to extract rent from knowledge. Current AI is, structurally, the same bet.
Beijing's Chip Ban Is Not Just Protectionism
China's rejection of NVIDIA and Western AI tooling is being read as nationalism. That is not wrong. But it is also a sovereignty play in exactly the same register as Current AI, just from the opposite direction. Both parties have correctly diagnosed that whoever owns the inference layer owns the cultural output. Moonshot AI's new Kimi model landing this week, already prompting breathless commentary about what commentators are calling full AI communism, is the cultural product of that sovereign stack strategy. The question neither side has answered is whether a commons can survive without a state behind it, and whether a state can build a genuine commons at all.