The usual way that people look at the agent life cycle follows a temporal order: pretrain, midtrain, posttrain. And then, almost as an afterthought, the harness comes. First you train the model, then you build the harness. However, in this post we will argue that the harness and pretraining, the two ends of the pipeline, are secretly holding hands. And this comes in two directions: ...
I had the privilege of attending ICLR 2026 this year thanks to generous support from UCB Sky Lab. I helped present Agentic Context Engineering at the main conference, as well as two oral presentations at the Lifelong Agent and MemAgent workshops. As a result, much of my experience centered on agent memory and continual learning. I am writing this reflection to summarize my takeaways from conversations with other researchers. I hope this blog strikes a chord with fellow readers and not with any companies’ legal teams over leaked information. ...
I had the priviledge to attend ICLR 2026 this year due to the generous support from UCB Sky Lab. I helped to present Agentic Context Engineering at the main conference plus two oral presentations Lifelong Agent and MemAgent workshops. So my experience mainly centered around agent memory and continual learning. I am writing this reflection mainly to summarize my takeways from conversations with other researchers. Hope this blog strikes a string with fellow readers and does not strike strings with companies’ legal teams for leaked information. ...