Every team has barrels and ammunition. Which one are you?

Spent the weekend re-reading Keith Rabois’s framework on this and I cannot stop thinking about it. Quick context if you have not come across it. Rabois has been talking about this since like 2014, his “How to Operate” talk at Stanford. The idea is simple.

In any company you have got two types of people.

Barrels are the rare ones. Point them at a hill, they figure out how to get over it. They take an idea from inception to shipped and pull people along with them.

Ammunition is everyone else. Talented. Valuable. Necessary. But they need a barrel pointing the way.

The stat that stopped me

PayPal at acquisition had 254 people in Mountain View. Even with that legendary team, only 12 to 17 were barrels. Roughly 6%. Rabois says that ratio is normal even at great companies.

I have read this framework probably four or five times over the years and nodded along each time. But this weekend it hit different. Because here is what AI just changed.

Ammunition got dramatically cheaper. Barrels did not.

Anyone with a laptop and Claude can produce ammunition-grade output now. The clean PR. The passing tests. The take-home that compiles. The artifact that used to be the signal of “this person can execute” is now just the signal of “this person has access to a model.”

The gap between barrels and ammunition just widened. Not closed.

Nobody’s interview process is built for this

Leetcode tests execution. System design tests execution at a different altitude. Take-homes test execution with a deadline. AI fluency tests are mostly leetcode with copilot turned on.

None of these things tell you if someone is a barrel.

A barrel and ammunition now produce identical output. They did not get there the same way. The difference is in the path. Did they push back when the spec was wrong. Did they notice when the model confidently solved the wrong problem. Did they ship something rough and iterate, or did they wait for someone to tell them what good looks like.

I do not think anyone has really cracked how to interview for this yet. But the teams that figure it out first are going to pull ahead in a way that compounds fast.

So back to the opening question. Which one are you on your current team? And more importantly, when was the last time you actually thought about it?