China Tech Talk

I’m staring a new series of what I call quick takes. These will be shorter (15-20 minutes) than scheduled podcast. They will be recorded whenever something interesting comes up in the tech world and I don’t want to wait to talk to someone about it.
So the first thing I did here is Kimi K3 and GLM-5.2 and what this tells us about China’s AI development. I think the overall story I want to put out there is the self improvement part of it and also models likely expanding over time and improved training data and flywheel with the Chinese labs.
By this point next year, we could very well (imo) see 5T+ parameter models coming out of China.


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tphuang.substack.com

This week, I had a chance to go on Carl Zha’s livestream to talk about the launch of Kimi K3 and what we saw at WAIC. Unfortunately, I had to leave early due to family obligations, but I think most of the issues I wanted to cover was already talked about. I’m actually a little surprised of the audio quality in there. I think next time, I have to go with something other than iPhone when I do this.
I just want again to point out something that I think is very under reported out there. Chinese models are getting larger and they will keep getting larger. Biren Tech reported this week that Chinese models will get to 5T+ range next year and it looks like we will reach 10T the year after as larger super node becomes available that can pool together all the high bandwidth memory and do efficient inference of these large MoE models. All of the major Chinese AI players like Moonshot, Zai, DeepSeek, Qwen and Minimax are going into new and larger sized models this year. Larger models require more compute to train, more data and more longer term optimization for inference. All of this has been made easier by the increased automation these firms can use to improve and speed up their own model development and training time. We are now seeing these very fast iterative improvement from Kimi and GLM in recent times. These are the new standards. The public is no longer happy with a new big improvement every 3 to 4 months. This needs to be every 1 to 2 months and eventually every month by the end of this year. Teams will need to have both flagship LLM and smaller and cheaper distilled models.


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tphuang.substack.com

This week, I invited Electric Felix on my show to talk more about EVs. I first encountered him on X/Twitter and then followed his journey on EVs in Europe as well as his trips in China and what he experienced there in Yunnan and charging stations. It was great talking to him and getting a sense of where Europe’s EV infrastructure and demand look like at the moment. I was in Spain and Italy earlier this year and the latter is still very hydrocarbon centric. We ended the talk with how the current trade dispute with Chinese EV industry in Europe gaining market share might get worked out. I think we will probably end up just seeing more localized production over time.


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tphuang.substack.com

I had the pleasure recently to go on Carl’s live stream to talk about the recent LM-10B launch and recovery. Why this was significant and why China picked its design of recovering in offshore pad. We also talked a little bit about the strategic implication of having low cost launches and what else China needs to do to continue to make advancement in space missions like building up LEO constellations and also do Moon landing and building lunar colony. We also had some fun talking about SpaceX/Elon part of this.
Finally, I did Q&A from the audience. A lot of AI related questions, especially with AGI stuff.


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tphuang.substack.com

I’ve been wanting to do an episode on GLM-5.2 & Zhipu for a while, but didn’t get a chance to do so. Last week, I did a live stream on Carl Zha’s channel and here is my piece on it.
I’ve had very good experience with GLM-5.2 so far and will probably move up a level in coding plan if their next release makes a similar level of jump. I’ve also had good experience with Kimi 2.6 & 2.7. Although in terms of Software engineering, GLM-5.2 is clearly better from my personal experience.
I would just recommend everyone to get off Opus and Claude Code. Try the cheaper GLM through ZCode or Kimi through Kimi Code. Or you can get both. Sometimes, one model would be better and other times, the other one will be better.


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tphuang.substack.com