There has been a lot of talk lately – not for the first time – that we might be losing control of AI and thus need to slow down development at the frontier. I’m still having a hard time figuring out what to make of these warnings, but it’s interesting to see how one narrative keeps coming up in tandem with them: «But China!» The argument goes like this: «It might be wise to slow down AI development, because terrible things could happen. But then China just takes the lead and even worse things will happen.» This is a helpful conversation that dissects this argument and helps to understand the complicated relationship the US and China have when it comes to developing AI.

From Weekly Filet #586, in September 2026.