Skip to content

China's approach to AI and robotics | Lex Fridman and Jensen Huang in Taiwan

Read the Portuguese translation →

Author: Lex Clips · Watch on YouTube

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: China’s approach to AI and robotics | Lex Fridman and Jensen Huang in Taiwan · Channel: Lex Clips · 01:10 · Watch on YouTube

Topics: Model inference and optimization, AI infrastructure and hardware, Markets and economy, Robotics and autonomous systems

Lex and Jensen discuss visits to AI and robotics companies in Shanghai and Hangzhou. They note that Chinese AI companies are short on computing resources, but say these constraints encourage efficient architectures and inventive solutions. The conversation closes with remarks about funding and IPOs, followed by thanks for the experience.

  • They visited MiniMax and other AI companies in Shanghai.
  • They also visited Unitree in Hangzhou and met robotics researchers.
  • The robotics people are described as humble.
  • AI companies say they are starved for computing resources.
  • The speakers suggest resource constraints can lead to efficient architectures.
  • They describe scarcity as a driver of invention.
  • Despite severe constraints, companies are still getting funded and making progress.
  • They mention that many IPOs are happening.
Start Chapter Summary
00:03 Visits to AI and robotics companies They discuss visiting AI companies in Shanghai and robotics company Unitree in Hangzhou.
00:36 Scarcity and invention They reflect on how limited computing resources can encourage efficient architectures and inventive solutions.
01:10 Closing The clip ends with a brief closing remark.

They’re compute starved. — 00:03

More compute. — 00:03

constraints is the kind of mother of invention. — 00:36

They’re still getting it done. — 00:36

MiniMax, Unitree, Escassez de capacidade computacional, Arquiteturas eficientes

00:03 Did you go to Shanghai? >> Yeah, for for some of the MiniMax, some of the AI company. >> Okay, yeah. Boy, those companies are good. >> And then we went to Hangzhou for Unitree. And those the robotics folks, full of humility. There’s a lot of robotics companies. And you know what all the AI companies are saying? They’re compute starved. >> Yeah. >> More compute. >> Yeah, but somehow somehow, you know, when they’re compute starved, it forces them to come up with these incredible architectures. They’re so efficient.

00:36 >> Yeah. Yeah, constraints is the kind of mother of invention. >> Right. >> Scarcity. >> Yeah, they’re severely severely constrained. And they’re still they’re getting funded. They’re still right. >> Mhm. Yeah. Many IPOs. >> Yeah. They’re still getting it done. >> I think so I got >> [laughter] >> All right, Jason. Thank you for an experience of a lifetime. That was that was delicious. >> Thank you. That was fun. That was incredible. >> was awesome. >> All right, we got to get it tomorrow.

01:10 >> Still going.