Keep pulling the thread on Kevin Xu.
Researchers at Western AI labs like OpenAI, Anthropic, DeepMind, xAI, and Meta have access to one or two orders of magnitude more compute than their counterparts in China.
The operating environment for technology companies in China is defined by US export controls, tech sanctions, and China's own domestic crackdowns.
Open source is one of the best, if not the only, strategies available for Chinese tech entrepreneurs to expand their businesses beyond China's borders.
The business environment for IT in China has historically been poor, as companies prefer hiring people to build custom solutions rather than paying for software.
Chinese software companies must expand internationally to find markets with a stronger willingness to pay for technology, such as the US, Western Europe, and Japan.
China's tech crackdown, particularly on education and gaming, released a large number of unemployed engineers into the market, which suppressed the nascent trend of companies paying for software solutions.
Lin Junyang, a tech leader of the Qwen model team, publicly resigned from Alibaba.
Mira Murati's new company, formed after she left OpenAI, is reportedly raising funds at a $50 billion valuation.
Yann LeCun's new AI lab raised $1 billion in its initial funding round.
Both OpenAI and Anthropic have published reports claiming to have detected Chinese state-affiliated hackers using their models for malicious activities.
The business upside for individual contributors and founders in the Western AI ecosystem is one or two orders of magnitude greater than in the Chinese AI ecosystem.
Chinese AI startups like Zhipu AI, MiniMax, and DeepSeek have not achieved a level of success comparable to OpenAI, Anthropic, or even tier-2 Western AI startups.