By noon yesterday I'd written most of an MVP for a demo, shipped a production hot fix for a meeting embeddings corner case, and sat through a customer call. By evening I couldn't tell you which of those actually mattered.

Most careers reward going deep in one domain. That was me at Apple for fourteen years. Founding a company, and doing it alone, is the opposite sport.

Any Given Day

Here's the last seven days. One person. Not cherry-picked.

UI/UX mockups. Product specs and architecture calls. Most of the code. Integration work and bespoke deployments. Marketing copy and a new demo video through Remotion. CI/CD through Github Actions. A custom CRM that's half Attrove, half markdown files. Customer research breakdown. Investor updates. This blog post.

That's the life of a solo founder with agents for coworkers. And most days it works.

The Hard Part

The adage of wearing multiple hats holds today just like it did pre-AI. But wearing them isn't the hard part now. Fifteen minutes, a well-written prompt, a little orchestration, and you're back kicking off agents to perform work.

The hard part is deciding what deserves your attention today. You can do anything, so capability isn't the constraint. Judgment is. Taste is. Knowing what not to build is.

Should I work on the demo? Definitely need to get a new video out. Found some backend bugs that need patching, though they aren't customer facing…yet. Agents will happily build, optimize, and deploy whatever I point them at. They don't care if it's the wrong thing.

The Human in the Loop

Chances are, if you're a salaried employee happily taking home a paycheck, someone else decided what mattered. A PM wrote a roadmap. A manager delegated a special assignment. Your job was to execute with precision, and ship on time.

Now the calendar stands as a reminder of what could be. What deadline do you set? What date do you aim for? I set a product roadmap two weeks ago and I've already taken issue with it.

Last weekend I spent three hours ensuring Attrove’s landing page is Agent ready (using Cloudflare’s new Is Your Site Agent-Ready? framework). It shipped. It’s near perfect. Nobody asked for it. Meanwhile, an unfinished customer demo has been sitting for three days.

It's so easy to keep building and shipping code because it feels like progress. Progress toward what, though? The ambiguous work of measuring traction, or figuring out if this customer is actually a fit. That's harder than running any agent.

The Good, the Bad, the Slop

The pace of change has accelerated tremendously. New ways of imagining work, new ways of shipping it. And honestly, building today is genuinely fun.

The flip side: it's so easy to build and ship that the sheer volume of generated content is absurd. Recent analysis suggests more than half of long-form LinkedIn posts are now likely AI-written. Over half. The feed you're probably reading this in. (side note: I choose to write my own pieces to sharpen my thoughts and express my feelings; I use AI to polish the outline and perform research but not the writing itself)

We're likely, for the first time, more bottlenecked by attention than by supply. Making something worth finding, and getting found, is where the work is. Distribution is brutal.

What's my focus going to be today? Sending my writing off into the ether? Check. Over-engineering my product? Most likely. What's the one thing worth doing? Now that's the hardest question to answer.

Let's Chat

Yes, I'm aware I'm ending a post about distribution by asking you to find me. Sometimes you just ask.

If you're building solo, thinking about it, or just want to compare notes, I'd rather hear from you than write another post. Find me at [email protected]. And if your agents could use actual context about what's happening across your tools, that's Attrove.

Graduates of Wharton Online's Real Estate Investing Certificate Program are already co-investing and building joint ventures together.

Use code SAVE300 for $300 off tuition. Starts June 8.

Learn More