Alice

Can we climb this mountain? I don’t know
Higher now that ever before

Amazing talk, the first hour of it is really great and aligned with my thoughts/understanding of agentic engineering https://m.youtube.com/watch?v=dHBEQ-Ryo24

As an engineer, I try to discount my thoughts on AI replacing human coders, after all, that’s my career. But this talk reinforced my views: with LLMs, the software we are building will be more complicated. So it’s not “you need less engineers to build more”, but “you need more engineers to build waaay more”. Because that’s what your competitors will do.

Amazing write-up by the legend Donald Knuth!

https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cycles.pdf

I’m also really looking forward to the future volume of The Art of Computer Programming. I have a really old edition from my dad (it traveled with me from the East Coast to the West Coast!).

I really enjoyed this read about RLMs https://www.dbreunig.com/2026/02/09/the-potential-of-rlms.html

Every new feature in agent harnesses I see is about effective context management (think agent skills, subagents, hooks, even your basic agent file with instructions). Seeing applied research in expanding context window is great!