The second seat

One board observer. Two coding-agent startups. One investor on both cap tables. At the end of September, Factory fired observer and adviser Chris Degnan, alleging he had discussed a job at its direct rival Cognition while sitting in on Factory's board-level discussions. Degnan said he had already resigned and denied passing anything along, then joined Cognition as chief revenue officer. Vinod Khosla, whose firm backs both companies, took Degnan's side.

This week Benchmark's partners told TechCrunch Disrupt that venture's old rule against funding head-on competitors can no longer be counted on. They're right about why, and the rule is going. But the Factory episode shows what the rule was standing in for. Founders never needed protection from a fund's other checks. They needed protection for what gets said in their boardroom. As colliding portfolios make the ownership rule unenforceable, the access rule is the one worth keeping, and keeping it should cost the fund something real.

Chetan Puttagunta gave the honest account. When building software gets close to free, ambitious founders want all the surface area they can see, and companies that started in different lanes run into each other. He drew the line where it belongs: backing a direct competitor at the outset is an easy no, while a portfolio company expanding into another's turf just happens. You can't decline a check you wrote two years before the overlap existed. OpenAI and Anthropic share dozens of investors by now.

Eric Vishria's reassurance is where it goes wrong. Benchmark keeps what it learns confidential, he said, and in any case there's little a partner could pass along:

That misplaces the risk. Nobody worries that a partner will smuggle a rival's architecture across town. A board deck is pricing, pipeline, burn, the customer about to churn, the executive about to be hired away. A rival would pay for every line of it, and reading those lines is exactly the expertise a board member is there to supply. The Factory allegation, whatever its merits, wasn't about code. It was about a person with a seat and an offer from the other side.

Peter Fenton supplied the better frame without drawing its conclusion. He placed investors on a spectrum: the co-founder at one end, who would never work on a competitor, and the public-market stock picker at the other, who can own several rivals because he only allocates capital. Benchmark, he said, sits near the co-founder end. Read it again and it's a spectrum of access. The stock picker can own both companies because he sees neither one's plan.

The rule that follows: when two portfolio companies collide, the fund keeps both positions but moves one of them toward the stock-picker end. One board seat, not two. No observer straddling the line. Information rights on the second company cut to what any shareholder gets. And the disclosure Fenton calls the golden rule, no surprise after the fact, made the day the overlap appears rather than the day it leaks.

A fund that wants co-founder access to both rivals is asking to be the co-founder and the stock picker at once.

The objection is that walls inside a small partnership are fiction. Partners share a Monday meeting, the fund's returns ride on both companies, and nobody believes a handful of people can unknow something. Concede it: an internal wall is thin. That's why the remedy has to be something the fund gives up, not something it promises. Surrendering the seat means losing influence over a company you own, which is the same cost the old rule imposed by forbidding the second check. The price moved. It didn't disappear. A fund that wants co-founder access to both rivals is asking to be the co-founder and the stock picker at once.

The second check is negotiable now. The second seat shouldn't be.