Okay, everybody is weighing in but I have a thing to say I haven’t seen elsewhere so I’ll just say it.

To recap, Paul Graham says that the orthodoxy for CEOs of “hire good people & let them do their jobs” is wrong. Founder Mode is the alternative—stay involved in all the details of the business. Micro-manage. Wander around. Talk to everybody. Circumvent all the channels. Correct the details.

First, I think Paul is essentially right but that the way he expresses his idea is shallow, ineffective, & prone to cause unneeded suffering. (Before I go further, I’ll point out that my experience is as one of the led, not one of the leaders. The times I’ve been the leader—patterns, JUnit, TDD, XP—my constraints & incentives were very different from those of commercial CEOs. Apply grains of salt.)

Purpose: Survival

First, the primary purpose of the kind of leadership Paul discusses is the continued existence of the business. Everything bows to survival. As I discussed previously, game over absorbs future winnings.

Survival is an asymmetrical game. There are many ways to die, a few ways to win. And those few ways are the result of long chains of not-dying decisions. What behavior leads to better chances of survival?

I’m going to speculate that deadly decisions are actually the interaction of two seemingly-innocuous decisions more often than they are one big blooper. [ed: need examples] If this is true, then the CEO really does have a unique role to play—spotting the Mentos & the Coke before they come together.

Here’s the thing—most decisions aren’t like this. Encouraging founders to get involved in all decisions (even if that isn’t what Paul said, that’s how some folks will read it) reduces the chance of catching interactions. Cultivating a sensitive sniffer for possible danger & only then getting involved increases the chance of survival.

Iteration

Second, business is an iterative game. CEOs don’t just call “Heads!” once & win or lose. Each decision has 2 outputs:

• The consequences of the decision itself.

• A new organization more or less prepared to make the next decision.

Paul ignores this second output entirely. What does it do to the people in the organization and the organization as a whole when the CEO jumps in and changes a decision in someone’s area of expertise? Even if the CEO is right, in the sense that the change increases the chances of survival?

Damage. Damage that needs to be repaired in order to restore an organization prepared to make good decisions next time.

Yes, there’s a benefit to the Founder Mode decision but there’s also a cost. Ignore the cost & CEOs will find themselves more likely to have to intervene at the same time as the number of decisions & the complexity of the interactions between decisions grows. Something will eventually give.

Chordal Connections

Founder CEOs have 3 (at least) advantages that give them a unique ability to create value (i.e. increase the survival chances).

• Social capital. The Founder CEO can escape local maxima that no one else would think of escaping.

• Breadth + depth. The Founder CEO knows a little bit about the whole business & the people involved. These chordal connections position them to spot and (using their social capital) avoid survival-threatening combinations of decisions.

• Incentives. Professional managers are also playing an iterative game, looking to their performance in this gig to set up their next gig. The Founder CEO is all in on this hand.

Wielding power comes at a cost. Ignoring that cost serves no one in the long run. We all love a good hero story, but it’s not the whole story.