i didn’t get why cursor made their own git system at first, but after checking out their agent swarm blog, it makes a lot of sense. highly recommend everyone to read it. they went from doing 1000 commits an hour to 1,000 per second in just a few months. that’s a 60x increase. crazy cursor.com/blog/agent-swarm-…

I open-sourced my AI-ready second brain & GEO/AEO setup. You get the structure, prompts, workflows and context files I’m using to turn notes into something agents can actually read, use and build on.

Introducing Headlong, an open source microharness for persistent agents: self-guided agents that think continuously. Most agent harnesses are reactive: you send a task, the agent completes it, and then it sits frozen until the next request. Cron jobs and heartbeats wake it up to run a checklist and put it back to sleep. A Headlong agent is never asleep. It keeps generating thoughts about whatever it decides is interesting, in a self-guided loop inspired by human inner monologue. Your message doesn't start a session. It's one more observation that lands in the agent's thought stream, and the agent decides if and when to reply. Headlong is built on the idea of persistent agency: continuous inner thought generation between external interactions. The agent sets its own interests and priorities, comes up with its own projects, and sometimes pings you unprompted with progress. To keep our prototype as simple and small as possible, we implemented Headlong as a microharness: a complete agent harness in under 10K lines of Bash, organized as a handful of small executables. It includes a loop that generates the next thought, shellm (a recursive language model written in Bash), a trajectory stored as a DAG of jsonl files, and context as a projection of that trajectory. We've been running one Headlong agent internally at Laude for several weeks. The whole team talks to it over Slack and Telegram, and every conversation lands in its single stream of thought. It works in its own fork of Headlong and we've pulled over 50 of its commits into main. One night, with nobody talking to it, it went back to check whether a recall process it had built was actually wired into its mind, found that it wasn't, diagnosed and fixed the bug, and verified the fix end to end. 48 minutes, no human asked for the fix or was in the loop at any point. Every step is a timestamped line in its log. Things broke too, and we wrote those up. Background thinking costs us $1 to $2 an hour, our agent stopped its own service three times by accident, and self-delegation died on day one. Details in the post. One line installs everything and starts an agent. Use a dedicated sandbox and spend-capped API key; it runs real shell commands and thinks around the clock. Headlong is research software, be careful! curl -fsSL headlong.ai/install.sh | bash Launch post: laude.org/updates/headlong-a… Repo: github.com/laude-institute/h… Headlong is a @LaudeInstitute / MIT collaboration.

More engineers should be working on company knowledge infrastructure. The opportunity for vast improvement across your teams is there to be taken. For ten years, company knowledge meant a wiki nobody updated and a founder explaining the same context for the fifteenth time, which was fine when humans were the consumers. Now agents consume company knowledge too they need it structured, governed, current, and shared. The companies I talk to aren’t struggling with model quality. They don’t care. They’re struggling with the fact that their sales agent and support agent quote different pricing, and nobody catches it for weeks until it’s a problem. One engineer working on the knowledge layer has more impact than five engineers building another agent. Because every agent is only as good as what it knows about the business, and right now most agents know almost nothing, and get confused about the little they do know - fast!

My X bookmarks used to be a graveyard. Now they automatically feed my second brain and become context for my agents. Feed this to your agent and clean up your own bookmark graveyard!

Human-in-the-loop is the easiest thing in the world to put on an architecture diagram. It is the hardest thing to make people actually do. I tried a review dashboard first. Didn't use it. Too much friction. Too far from where I already spend my time. Now the system sends proposed knowledge to Telegram. I approve, reject, or amend it there while I'm doing other things. Takes a few seconds per item. Approval systems fail when they feel like a second job! Keep it inline with your workflows.