Wasserman's data shows that most startups die from founder-team dysfunction, not bad markets. If you're wiring agents into your org before you've wired the humans, the dysfunction just moves faster.
Noam Wasserman spent years mining founder histories to find where the wreckage actually started, and the answer kept being the same: the first handshakes. Who gets equity and why, whether the split was negotiated or just defaulted to equal, who holds which role title at month two — these decisions calcify fast. Equal equity feels fair at the kitchen table and turns punishing the moment performance diverges. The rich-versus-king tension shows up early too: founders who optimize for control consistently sacrifice valuation, and those who optimize for valuation consistently lose the wheel. Neither is wrong, but confusing the two with your cofounders is.
When agents handle execution, the remaining human decisions are almost entirely judgment calls — and judgment calls require clear authority. Wasserman's core finding sharpens here: if your cofounder agreements are ambiguous about who owns what, an agent-native operation doesn't reduce the friction, it surfaces it on every deployment decision, every model swap, every customer escalation the agent kicks upstairs. The humans holding judgment need to know whose judgment wins. Sort the equity, the roles, and the rich-versus-king question before you hand anything to an agent, because the agents will find every gap.
- Default equal splits are a timer, not a gesture of fairness — negotiate contribution explicitly from day one
- the rich-versus-king question must be answered between cofounders before it gets answered for you by a term sheet or a product crisis
- agents accelerate every human-coordination failure already latent in your founding structure.
