unknown

Palantir CTO @ssankar on Deepseek:

“China is the best at long term planning— the one thing not in their plan, was AI. It is an American phenomenon.”

“It’s not a coincidence they had to steal Deepseek. Distilled from OpenAI. They literally had to steal that.” https://t.co/NXyr0G6QrT

kache

A provocative claim asserting that DeepSeek, a Chinese AI lab, was the first to develop test-time compute via reinforcement learning-based reasoning training, predating similar techniques attributed to American AI companies. The post frames this as an underreported fact obscured by narratives around model distillation, pointing to DeepSeek-Zero as evidence of priority.

Fyi, deepseek invented the technique of test time compute, i.e. reasoning trained through reinforcement learning. They were first. China was first. No amount of American propaganda aboutuh distillation should make you think otherwise. Look up the date of deepseek zero https://t.co/myDG7K6o36

<p>Fyi, deepseek invented the technique of test time compute, i.e. reasoning trained through reinforcement learning. They were first. China was first. No amount of American propaganda aboutuh distillation should make you think otherwise. Look up the date of deepseek zero https://t.co/myDG7K6o36</p>