Results
3 articlesSort
1:48:46ZH → English
Zhang Xiaojun Podcast· Nov 18, 2025Podcast · 2 speakersLiu Xianming: Removing Language to Bring Physical AI Back to Data and ActionGuest 刘先明 · Host 张小珺Liu Xianming, head of autonomous driving at XPeng, explains why language supervision in VLA training can become an efficiency bottleneck and how he is helping XPeng shift from an autonomous-driving team to a physical AI company. The conversation also covers scaling data, organizational change, production safety, and what comes next for autonomous driving.
1:27:02ZH → EnglishZhang Xiaojun Podcast· Sep 5, 2024Podcast · 3 speakersGuangmi: Self-Play RL Could Open a New Scaling Law for AGIGuest 广密 · Host 张小珺Guangmi believes the marginal gains from scaling language pretraining alone may be fading, making self-play reinforcement learning a more promising path. Whether it generalizes to coding, video, robotics, and beyond may determine the next stage of the AGI race.
1:33:07ZH → EnglishZhang Xiaojun Podcast· Aug 17, 2024Podcast · 2 speakersLi Guangmi on a Year in Global LLMs: Scaling Law, the AGI Bet, and the US–China LandscapeGuest 李广密 · Host 张小珺Drawing on his front-row view from Silicon Valley, Shixiang Capital founder Li Guangmi looks back at the competition among OpenAI, Anthropic, Google, and Chinese model companies. He argues that model capability and cost together will shape the next wave of applications, while today’s enormous investments are bets on whether intelligence can keep scaling and how AI will reshape the economy and industry over the long term.