TL;DR of my new article: I cut my agent's memory from 218 files to 6. The loud advice is "build a bigger memory system." I went the other way. Your memory and your instruction file (CLAUDE.md, AGENTS.md, whatever your agent loads) are the same problem: standing instructions that load every session and rot the moment you stop looking. 🧠 A fat memory makes your agent dumber. It injects every session, and a full context window performs worse. Keep each file under 200 lines 🗑️ Most "memory" is misfiled: trash (git already has it) or a skill lesson sitting in the wrong place 🛠️ Anything you do twice, write a skill. A lesson belongs in the skill as a PR, where it helps everyone, not journaled in memory where it rots 💾 Back up the whole store first, then cut to the bone. Deletion is only fearless when it's reversible 🔌 Turn auto-memory off. Move your survivors into the always-on instruction file so they still load 📄 Project instructions are one line: import the shared file. Write conventions once, every agent reads them (AGENTS.md is the cross-tool standard) ✂️ Stop using your instruction file like a trash can. Delete anything the model can infer from the repo. Every line earns its slot or gets cut 🔎 Still want recall? That's pull memory, not push. A brain you query, not one you carry into every session (gbrain, supermemory) Your instruction file is the most expensive real estate your agent reads. Pay for signal, not square footage.

Made a friend from college a "How to learn agentic engineering in 3 days" guide custom for him. I took our 30 minute call Granola transcript copy/paste of his linkedin copy/paste of my Every Agentic Hack I Know article and then /ce-plan make a vercel web page with check boxes of his 3 days of training. My fav part? It added a "stuck" button at the bottom of every page. I had no idea what it did. Text me? Nope - it said "take a screenshot of your problem and paste it into your LLM"

Fun article in Bloomberg today I was featured in. “‘It’s OK, Matt,’” he says with a laugh. “They said that agents were supposed to do our work for us, but I’ve never worked harder in my life. I just have 100 times the output that I had before.’”

The worst agentic engineering feeling in the world. What on earth was going on in these terminal windows before my computer crashed?!

Okay I'm building something. Hope it's good! "All 8 units now Full crash recovery native resume breadcrumb offer in one branch. Largest diff, in a ~60-open-PR hot zone, and consumes resume-binding APIs that #6631 is actively reshaping (higher rework risk)."