Episode
Why it Sucks to Work in AI in China + Open Source with Kevin Xu
- Podcast
- ChinaTalk
- Published
- Mar 12, 2026
- Duration seconds
- 3034
- Processing state
processed- Canonical source
- https://pscrb.fm/rss/p/traffic.megaphone.fm/CHTAL2856791081.mp3
Actions
POST https://stenobird.com/v1/public/podcasts/chinatalk-418044/episodes/why-it-sucks-to-work-in-ai-in-china-open-source-with-kevin-xu/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/chinatalk-418044/why-it-sucks-to-work-in-ai-in-china-open-source-with-kevin-xu.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Kevin Xu of http://interconnected.blog/ and I did a liveshow on substack! We chat about why working in Chinese AI looks so much tougher than building in the West: less compute, lower upside, more political constraints, and a much weaker market for enterprise software. We also get into Kevin Xu's definitive history of open source in China (https://interconnected.blog/chinese-open-source-a-definitive-history/?ref=kevin-xus-interconnected-newsletter) and talk why open source has become one of the few real paths Chinese AI companies have to win users abroad, even as the business model at home remains brutal. Also: the Qwen shakeup at Alibaba, what it says about the limits of China’s AI lab ecosystem, and why Chinese firms may still beat the West in areas like AI shopping and commerce. Learn more about your ad choices. Visit megaphone.fm/adchoices