Why do you think bigger teams are becoming a liability in the age of AI?

Every team I've worked in, big or small, had the same complaint: we don't have enough resources to get everything done. Well, now we have it. With capable AI models, everyone on a team (designer, developer, PM) can get a lot more done. But in big teams, those new capabilities smack head first into the decision-making and alignment problems that have always been there. In small teams, they don't.

That's really the crux of it. The tools were never the bottleneck. All those processes in larger organizations (review, feedback, present to leadership, go back, repeat) exist to produce decisions and alignment. And they often fail because people in big companies don't want to make decisions, plus everyone wants to leave their mark on things. So when AI lets each person produce 10x more, you just get 10x more stuff waiting in the same slow approval line.

I've seen the small team version firsthand. When Twitter acquired Bagcheck in 2011, Sam (our CTO) and I were shipping multiple times a day. With just two people building the product, we never got bottlenecked on decision-making or coordination. Bagcheck never crossed four employees, and when Google acquired Polar in 2014 there were five of us. Neither company had AI coding agents. Today, the number of people you need to get a lot done fast is probably a lot smaller than you think.

There's also a cultural cost that big orgs have built up. For years, most designers were effectively designing for approval: get it signed off by product, engineering, VPs, legal. Along the way the muscle for pushing a strong point of view atrophied. And if your job is mostly following precise instructions, whoever writes those instructions can now hand them to an agent.

Of course, I don't think team dynamics are worthless. Reasoning is fundamentally social, and the tension between designers, developers, and PMs can drive better outcomes. But the cost of generating ideas is moving toward zero while our ability to evaluate and synthesize them stays capped. So you want enough people to sharpen each other's judgment, and no more people than it takes to make a decision.

Put simply, AI productivity gains hit walls of decision making and alignment, so once again small teams move faster, way faster. If you're stuck in a bigger org, the closest thing to a fix I've found is to pin your key decisions and point of view in one place, so you're managing one thing instead of a hundred.