The Cluster

Solo founders are winning. The data's clear. But winning alone is still alone.

Solo founders are 2.6x more likely to own an ongoing, for-profit venture than teams of three or more. The share of startups with solo founders and no VC has climbed from 22% in 2015 to 38% in 2024. And 60% of ventures that fail do so because of team issues—conflict, equity disputes, slow decision-making.

Sam Altman is betting on one-person billion-dollar companies. I get why that's exciting. AI agents are replacing entire workflows. Solo founders are outperforming teams. But I think they're asking the wrong question.

The unicorn is a symptom. The real story is who's running the experiment—and what they actually need.

A-players are leaving

Right now, a new type of founder is emerging. A-players who could get any job, raise any round, hire any team—but chose not to. AI-first. No employees. Full ownership of their time, decisions, and upside.

They're not chasing a valuation. They're testing how far one person plus AI can actually go.

The old model assumed that building big required building a team. That assumption is dead. AI removes the execution bottleneck. Now the only limit is what you can see.

The loneliness problem

Here's the paradox. Even if solo founders are winning, the journey is lonely.

I've built companies for 20+ years. Multiple times I've started from zero, scaled, failed, started again. And the loneliest moments weren't when things were hard. They were when things were working and I had no one at my level to think with.

Not employees—they look to you for answers. Not investors—they want updates, not raw thinking. Not friends outside the game—they can't feel the weight of the decisions.

What worked best for me wasn't a co-founder. It was other A-players building their own things, who understood the game, who could call bullshit on my thinking, and who I could do the same for.

Even the best founders need peers.

The anti-community thesis

The obvious answer is a community. But I've grown skeptical of communities.

Most communities are where ambition goes to get diluted.

Slack channels become graveyards. Weekly calls become obligations. Content becomes noise. There's a community manager trying to keep people engaged, metrics being tracked, gamification being added.

Communities optimize for the creator of the community. They grow by lowering the bar. The more members, the more impressive the numbers look. But that's exactly backwards for A-players.

A-players optimize for signal. They don't have time for noise. They don't need engagement tricks. They're already engaged—with their own work.

What I'm thinking about instead is a cluster.

No structure. Just collisions.

This idea is still early, but here's how I'm thinking about it.

A cluster is different from a community in a few important ways.

No structure. No programming. No community manager manufacturing interactions. No content calendar. Just proximity.

Things emerge from clusters. Group chats form organically—the kind that actually replace your need for a team. Four or five people who think at the same level, building different things, but trusting each other enough to share real problems in real time.

Partnerships happen because you already know each other's work. Retreats happen because you actually want to meet, not because someone scheduled a networking event.

The difference sounds subtle but it's fundamental. One optimizes for retention metrics. The other optimizes for quality of collision.

Clusters are self-policing. If you're not useful, you drift out naturally. No drama. No awkward removal. No rules to enforce. The only filter is whether people want to keep talking to you.

The answer isn't another platform. It isn't another community. It's just collision. Repeated, high-quality collision between people who are actually building.

Finding each other

I'm looking to connect with founders building this way.

Not a specific number. Not a program. Just finding the right people and seeing what emerges.

The profile:

• Building a startup with zero employees—by choice, not circumstance
• AI-first. Not just using ChatGPT for emails, but building the company around what AI makes possible.
• A-players. The kind of person who was always the engine wherever they worked.

Maybe it starts with X for daily signal. Maybe small group chats form. Maybe a retreat once a year. I don't know yet.

The point isn't to build a community. It's to create the conditions for a cluster to emerge.

If you were always the engine

The idea underneath all of this is that the future of startups isn't companies. It's coalitions. Small, high-trust networks of builders plus AI, working in parallel, helping each other without merging.

I don't know exactly what emerges. That's the point. Clusters aren't designed. They're conditions for emergence.

But I know the type of person I want to find. Someone who was always the engine wherever they worked. The one who shipped when others talked. The one everyone relied on. Now they've stopped making other people's dreams happen and started making their own.

If that's you, I want to meet you.

Not a pitch. Just: who are you, what are you building, what's the experiment you're running?

DM me on X. Let's see what emerges.