China Solved AI Without Our Training Data. Now They're Writing The Rules
We Thought We Were Winning. We Already Lost.
On my flight from Nashville to Boston after Autodesk University 2025, where I'd just spent three days discussing AI and infrastructure, I couldn't sleep. Not because of turbulence, but because of what DeepSeek achieved with pure reinforcement learning - no Western training data needed.
By the time we hit cruising altitude at 11:47 PM, I had three documen…
Keep reading with a 7-day free trial
Subscribe to The Pacific Divide to keep reading this post and get 7 days of free access to the full post archives.