Recently, while listening to a podcast interview featuring Ben Horowitz, I was struck by an analogy he used to describe the bind founders find themselves in when hiring execs for roles they have never themselves occupied. “It’s like trying to hire a Japanese interpreter when you don’t know Japanese,” he said.

That hire is not only a decision you make, it is a decision you quietly decline to make: the decision not to do the difficult, slow work of learning Japanese – at least, enough to assess the person you are hiring from a position of knowledge. The bill for that non-decision may not arrive for a year or more. But it will come due eventually.

The interpreter line was a small moment in a much larger argument Horowitz was making, in conversation with Sequoia’s Brian Halligan, about what separates great founders from everyone else. Most founders do not fail from ineptitude; rather, they fail from avoidance. Horowitz outlines a few recurring patterns — decision debt, the VP of Sales trap, the long habit of running from the truth to preserve someone’s feelings — all of which are expressions of the same underlying behavior. The hard call is available. The founder declines to make it. And the cost compounds.

This avoidance is a sort of psychological minimum payment.

Anyone who has carried a credit card balance understands the mechanics. The minimum payment covers your bases, but leaves the principal almost entirely intact while the interest does its work. Decision debt functions in the same way. The principal is the hard decision you owe, the one you keep deferring because making it would require admitting something you would rather not admit. Avoidance is the minimum payment that keeps the account looking current. And the interest, which is the part nobody budgets for, comes due in lost optionality, in talent compounding in the wrong direction, or in a market that moves while you are standing still.

I have written before about the founder-investor bond as a kind of marriage (eight years on average) and about how the partnerships that endure are the ones where neither party swallows their tongue. Skipping the awkward conversation about the underperforming hire, softening the board update, telling your co-founder what they want to hear: each of these decisions preserves harmony in the short-term, and all accrue interest in the long-term. You have simply moved the cost to a future in which it will be larger, the way an unspoken resentment between two people does not dissolve but compounds, surfacing years later in a form neither of them can quite trace back to its origin.

The interpreter hire is decision debt incurred at the moment of hiring, and it is the cleanest example of how the interest compounds rather than staying fixed. Senior leaders hired without proper evaluation do not simply underperform in isolation. They hire a team in their own image, set a strategy you are even less equipped to second-guess, and establish norms that calcify long before you understand what you bought. This is Horowitz’s VP of Sales trap precisely (see our previous blog post on this topic): the wrong senior hire is not one bad decision but the first in a chain of decisions you have now outsourced to the wrong person. By the time the truth is undeniable, the principal you owe has grown into something far larger than the call you avoided.

The same pattern explains a phenomenon I see constantly in the companies we look at, and it is the one I find most clarifying in this AI moment. We talk about technical debt as though it were the disease. In fact, it is only a symptom. The legacy stack that needs an ever-larger engineering team simply to stay in place, that cannot absorb new information without months of work, that quietly dictates what the company is allowed to attempt next? That stack persists because, at some point, someone opted out of making a decision. The switching costs appeared too high, the rebuild too painful. So the company made the minimum payment, again and again, and called it prudence.

What the current cycle has made vivid is the magnitude of the debt that has accrued. Many of the most capable SaaS companies were architected before LLMs were a practical building block, and they are now competing against companies that were not. The pre-LLM incumbent needs many more engineers to run in place while a leaner competitor, built natively for this era, adapts at a fraction of the cost. The firms that chose to rip and replace, that absorbed the painful truth early rather than financing it, did not end up with less technical debt. They ended up with less decision debt. That is the advantage, and in a moment when capital efficiency is suddenly achievable in ways it never was before, it is becoming the whole game.

As any investor who has sat on boards of venture-backed companies will recognize, the instinct to deliver good news and bury bad news is not a personal failing; it is a gravitational force, and it runs downhill. A board that is managed rather than informed teaches the CEO to manage rather than inform, and the CEO teaches the same lesson to the executive team, and so on, until the entire organization is fluent in the minimum payment. Every good-news-only meeting is a payment that touches no principal at all.

A different relationship to the truth is required in order to actually retire the debt. The founders I most want to back are not the ones who avoid hard decisions gracefully. They are the ones who have made surfacing uncomfortable information a discipline rather than a response to crisis, the ones comfortable asking hard questions on a normal Tuesday, and not only in an emergency. Which decision have we deferred for more than a quarter? Where are we preserving feelings at the expense of facts? Who did we hire because we could not differentiate confidence and competence?

The interest on those questions never stops compounding. The only thing you control is whether you pay down the principal. As in any relationship worth keeping, the bill comes either way. The only choice is whether you settle it now, or later.