DeepSeek's claimed $15 million training cost was inaccurate, and open-source AI economics are more complex than advertised.

The US open-source AI model Thinking Machines was distilled from the Chinese model Kimi K2.5, meaning American open-source innovation is now building on Chinese foundations.

“It was distilled off a Chinese model, Kim K 2.5.”

AI model training costs must be recovered within 12-24 months due to rapid model obsolescence.

“You got to recover it pretty damn quick because it only lasts, you know, 12 24 months before it's obsolete.”

GLM 5.2 contains watermarks from another model called Mythos, indicating it was distilled from that model.

“GL GLM 5.2 has watermarks from mythos all over it.”

DeepSeek's claimed $15 million training cost to build an OpenAI competitor was not accurate.

“What did Deep Seek claim their training model was back when it launched? 15 million to train an open AI competitor. Like of course it wasn't true.”