If a tree falls in the woods...
I have now shipped several apps this month using AI only. No meetings. No sprints. No backlog. No Scrum Master. No Six Sigma PMs. I described what I wanted in plain English and the AI wrote the code. That part was easy.
The hard part is what I'm asking myself every time I ship one: does anyone care?
The people I know who are paying attention are shipping more software in six months than they did in the previous six years. A lot of CTOs I know — people who hadn't written a line of code in a decade — are back at the keyboard. It's fun. It's fast. It's cheap. And it's not clear anyone on the outside notices.
And now you can build from anywhere.
Cursor just dropped an iPhone app. You can vibe-code from the airport, from your kid's practice, from the line at Trader Joe's. Grok has it. Codex has it. I've been doing it with Claude Code for months.
The barrier to creating working software just dropped to "having a phone and a thought." Which means the barrier is no longer the barrier. Something else is.
So what's the actual problem?
Nobody sees what you shipped.
That's it. That's the whole thing.
Five years ago, having a working product was the moat. Building it was rare. It stood out. Now I can describe a product in a sentence and have it built by dinner. So can you. So can ten thousand strangers who haven't even started yet.
What's rare in 2026 — the thing nobody can vibe-code into existence — is distribution. An audience that knows your name. A list that opens your emails. A community that notices when you ship. The trust that makes someone hand you their inbox, their calendar, or their credit card.
That's the moat now. Not the build. The reach. Having someone who cares.
What I'm doing about it
I write this newsletter every week. I post on LinkedIn. I post on X. I'm building TextMyAgent in public — the homepage, the blog, the FAQ, every word indexed for search, every page pointing at one phone number and one product.
I'm playing the distribution game as hard as I'm playing the build game, because the build game is worthless without people using what you built.
The hard part is convincing one person to care, then ten, then a thousand. The hard part is the years of compounding before anyone notices. The hard part is choosing to be in 200 rooms instead of 1.
The greatest challenge ahead is obscurity
Amazing things will be built with AI in the next 24 months. Probably 90% of them will be ignored.
The 10% that aren't will belong to the people who started building distribution two or three years ago.
If you haven't started, today is the day. Pick the channel that fits your voice. Show up there weekly. Pick a name. Pick a niche. Start being seen.
The code is free now. The audience isn't.
What I'm watching
- Cursor mobile — Vibe-code from anywhere. The line between "I had an idea" and "I shipped a prototype" is now 20 minutes on the couch.
- OpenAI Codex — The coding agent baked into ChatGPT. Same trajectory, different surface area.
- xAI's Grok Code — Four full-throated dev-agent products shipping at once tells you where the floor is headed.
- Claude Code — What I personally use to ship most of what I ship. The quiet one that does the work.
- Fable Studio — Their AI Showrunner is opening up wider. AI-generated TV shows are moving from lab demo to consumer product. The next few months are going to be interesting.
- Google × SpaceX — Big-tech infrastructure keeps consolidating. Every AI product now depends on decisions made by two or three companies you don't get to vote on. Worth watching who's picking sides this quarter.
What I'm building in my own 'lab'
I'm REALLY pushing TextMyAgent — a personal AI agent that runs your email and calendar from your text messages.
I built it for myself because I got tired of the back-and-forth scheduling and the noise of email triage. $99/year, 14 days free, no card to start. If you've been waiting for a way to actually use an AI agent in your real life, this is the easiest on-ramp I know.
Also doubling down on the Essentialist platform — the agent engine behind this newsletter,and behind a dozen other distribution plays I'm running. If you're a founder who knows you need distribution but doesn't want to manage 17 more tools to do it, we should talk.
Need something? Just reply. CTO consulting at $500/hr. Dev team for hire. Media services. Or just punch back on this essay — I read every reply.
— Keith