I started this blog in 2003, and the first few years were genuinely exciting.
This was before social media, when blogs still felt like a place. I would write something, someone else would respond on their blog, and then a third person would join from theirs. We linked to one another constantly. Comments and trackbacks pulled the conversation together. RSS made it possible to keep up with people without all of us having to be on the same site. I wrote that year that blogs were a better model for building community than mailing lists. They were, at least for me.
People read what I wrote. More importantly, they replied. They corrected me, disagreed with me, added things I had missed and quite often took the idea somewhere better. On the blog's fourth anniversary, I wrote that "most of the motivation for writing here is the feedback loop." That feedback was part of how I worked: commit an idea to writing, let other people challenge it, and keep going.
I joined Twitter in February 2007. A month later I wrote "I Don't Get Twitter", which held up for about three months. By June I was using it and writing about what it could become. A lot of the people I knew through blogs were there, and for a while it preserved some of the same feeling. The responses came faster. The ideas got shorter.
By 2010, I could see what Twitter had done to my writing. I had written almost 1,100 posts over seven and a half years, but more and more of the things I might once have worked through on the blog ended up as short, temporary updates. I announced that I was going to fix this by publishing one substantial article every month. I managed two. (Apparently a content calendar was not the missing ingredient.)
LinkedIn followed a different path. When I joined, it wasn't social. It was a profile. You filled in your work history, connected to people you knew and mostly let the page sit there. Eventually LinkedIn became a place to publish too, but the writing there always felt performative to me. It wasn't really discourse or conversation. It was "look at me." (And, to be fair, I did plenty of it.)
Through all of this, the way I worked didn't change much. I tend to commit to an idea hard enough to test it seriously, while staying open to where the evidence leads. I let things incubate. I build experiments. I pivot when something breaks, and I keep looking for the leverage point that makes the rest of the problem move.
Writing was part of that process. It gave me an outlet for the thoughts that needed more than a few lines and made an idea concrete enough to challenge. The responses showed me where the argument was weak, where I had missed something, or where there was a more interesting idea hiding inside the first one.
Over the past year, that process has had me writing regularly again. Most of it lived in prompts, plans, specs and long explanations of what I wanted an agent to build. The work was better when I explained the goal, the constraints and why I cared about the result. When I didn't, the agent built something adjacent to what I meant, and I got to read my own ambiguity back in code. (This remains an annoyingly effective editing service.)
Agents fit that way of working unusually well. I can put down a half-formed idea and get something back while I am still working through it. Sometimes the feedback is direct. Sometimes the agent misunderstands me. Sometimes it does exactly what I asked and produces the wrong thing, which is a particularly efficient way to discover that the instructions were not the problem. I revise the prompt, but I also revise the thought behind it.
Working with agents rewards the kind of writing I had stopped doing. I can explain the whole thought, follow the side paths, supply the history and keep going until the idea is actually clear. I don't have to compress the thought to make it travel. Going deeper and being clearer actually improves the work.
The people who read the old blog regularly accumulated context about me over time. They knew what I had worked on, which arguments I kept returning to, what I found interesting and which tradeoffs I was likely to make. Agents need that kind of orientation too. A prompt can describe the task in front of us. It rarely explains enough about why I care, what I value or what I mean by good.
I've found that agents do better work when I give them a purpose, an identity and a frame of reference. Prompts, skills and conversations explain the immediate work. Longer-form writing gives them the context around it: how I arrived at the idea, what else it connects to and why one tradeoff feels right to me when another does not.
Writing that context is useful for me too. A plan for an agent system turns into a thought about managing people. A question about context takes me back to bounded contexts in software architecture. Something my law school advisor told me in 2004 describes work I am doing in 2026. Ideas I have carried separately for years begin to crystallize because I have to explain how they connect.
The blog gave me somewhere to put a whole thought, and the feedback gave me a reason to keep sharpening it. Agents make that loop active again. The response no longer depends on squeezing the idea into a feed or waiting to see whether an audience finds it. It starts while I am still writing.
That is why I am writing publicly again. I still want people to read and respond. That was always part of the joy, and I hope it becomes part of this again. The agents keep the feedback loop moving in the meantime, giving me somewhere to test an idea, refine it and find the parts I haven't really figured out yet.
The next things I want to write about start with the projects themselves: what I built, which assumptions I tested, what the agents made possible and which failures changed the next move. The systems have names, but the ideas should make sense before the names do. I want to ground the writing in the actual work because that is where the useful parts came from.
I won't promise a cadence (the archive has already heard that one). But I have the outlet back.