The Director Model: Why Maker/Manager is Obsolete

Paul Graham's maker/manager schedule is dead.

Graham's 2009 gospel divided us into two species: makers who need 4-hour blocks of deep work, and managers who live in calendar confetti. Pick your side. Optimize accordingly.

That binary made sense when human attention was the bottleneck. When you could code OR coordinate, never both.

Here's what changed: AI agents don't need coffee breaks.

I can maintain 15 parallel workstreams while sitting in a compliance meeting. Review smart contract architecture in one window, guide UI iterations in another, architect system integrations in a third. Each AI agent holds the entire context. Each executes at 300x human speed.

This isn't delegation. When you delegate, you transfer ownership. This is direction—I'm multiplying my presence across parallel threads while maintaining creative control.

Think about what this means for financial services. We've been organizing teams around the maker/manager divide since forever. Developers in their caves. Executives in their conference rooms. What happens when that boundary dissolves?

The math is staggering. A senior developer refactors 100 lines in an hour. AI does it in 12 seconds. But here's the kicker—it's not about the speed. It's about what you do with that speed.

You don't just code faster. You spawn 100 variations, test them against each other, synthesize the best elements. You don't just design one system. You architect entire ecosystems in parallel.

The new role isn't maker or manager. It's director.

Directors don't touch materials or manage people. They orchestrate systems. They work through AI agents like a conductor through an orchestra—except each musician can play a thousand instruments simultaneously.

Remember when banks started giving away toasters? That's what we're doing with the maker/manager framework. Optimizing yesterday's constraints while the game fundamentally changed.

The bottleneck shifted. It's not execution anymore—that's essentially free. It's not coordination—AI handles the handoffs. The bottleneck is imagination. Vision. Knowing what should exist.

When you can build anything instantly, what's worth building?

This is the question keeping me up at night. Not how to be more productive—that problem is solved. But what to do with infinite productivity.

We're not getting better tools. We're getting collaborative partners that multiply our presence. The constraint isn't time or attention anymore. It's judgment.

The directors who win won't be the ones who type faster or meet better. They'll be the ones who can maintain coherent vision across a hundred parallel executions. Who can recognize quality when they can generate infinite variations.

Who can answer the only question that matters: what future should we build?

The infrastructure is ready. The mental model isn't.

Stop optimizing your calendar. Start orchestrating systems.