Archive
2026
5 posts from this year
When Models and Agents Evolve Together: Notes from Fuli Luo
Notes on Fuli Luo's interview about OpenClaw, agent training, long-context model design, and how research teams learn as their tools change.
Building the Systems Behind AI: Notes from Heng Liao
Notes on Heng Liao's discussion of Ascend, the eighteen-level pagoda, chip architecture, memory and interconnects, open source, and the discipline of building useful systems.
How Robots Learn to Get Better: Notes from Liyiming Ke
My reflections on Liyiming Ke's discussion of experience-driven robot learning, generalization, evaluation, embodiment, and the human questions raised by increasingly capable AI.
When Intelligence Becomes Infrastructure: Notes from Ming Zeng
My reflections on Ming Zeng's discussion of strategic uncertainty, AI industry evolution, AI-native organizations, and the changing value of human work.
Call for Abstracts: AGU26 Session NH052
We invite contributions on reliable and scalable geohazard intelligence, from multiscale sensing and geospatial AI to open data foundations.