Destroy All Contractors

Founders learn to distrust dev shops for good reasons — short-term focus, cash-burning, the people who couldn’t hack it elsewhere. So here is the opposite case: a small, senior, permanent crew is the answer to all of it, and the AI stack has only made that answer stronger. If you’re an experienced operator building a new product, with finite runway and no appetite for a wrong hire, this is the model I’d point you to.

The on-demand CTO isn’t a budget argument. It’s arithmetic.

You don’t need a full-time CTO, a VP of Engineering, a tech lead, and a couple of developers to do real work. You need one or two motivated developers on a roughly 25-hour week and one person — me — who plays CTO, tech lead, and, most important over the long run, engineering recruiter.

Agentic engineering is what turns that from a contrarian claim into arithmetic. A small, senior crew doesn’t just cost less than a large one — it out-produces it. The right workflow yields something like a 10x speed improvement, and it only improves quality in experienced hands. So the case for a lean, senior team isn’t a budget argument. It’s a quality argument. The bloated org isn’t just expensive. It’s slower and worse.

The economic efficiency is real: a few motivated people, a total monthly budget of $8–15K, and the leverage underneath them getting dramatically better every quarter.

Recruiting and arrangement is the whole game

Phil Jackson, Quincy Jones, Miles Davis — the skill that made each of them a legend wasn’t the thing they were famous for. It was recruiting the right people and arranging them into something unstoppable.

That’s the entire model. We assemble elite, low-volume, high-reputation crews tailored to a mission — Hollywood, not Uber. No race-to-the-bottom freelance marketplace. I don’t bring a random developer to your project; I bring people I’ve worked with for years, each with the same three traits: entrepreneurial instinct, attention to unspoken detail, and technical mastery. Talent is the moat — more so now than ever, because agentic tools widen the gap between a senior operator and a first-timer. The tools don’t flatten talent. They amplify it.

The Norman doors are back, at scale

The entire purpose of design is to not build Norman doors — the door, named for Don Norman, that gives you no honest signal about whether to push or pull. Integrated design thinking is inquiry, observation, and logical completeness, so the thing you ship doesn’t quietly confuse the person using it.

That has never mattered more. The market is racing toward a hundred-to-one leverage — agents overseeing agents, more output faster — and when you look closely at what that race actually produces, a great deal of it is thin UX and broken information hierarchy. Norman doors, generated at industrial scale. More code is not the achievement; more code is more entropy, more risk, more surface for a customer to bounce off of. Less code is the feature. Making the tool honest about itself is exactly the discipline the hundred-to-one race abandons.

That’s the whole reason I run on specify, run, evaluate. The mastery lives in the specification and the evaluation, not in the volume of what gets generated. Every unit of output is a unit someone eventually has to judge. A team optimizing purely for speed is manufacturing doors nobody can read.

Longevity is the actual product

Across our founder relationships the average founder stays with us about four and a half years, and that’s a floor, not a ceiling, because most of those relationships are still active and still compounding. Our oldest dates to 2016 — roughly a decade. That founder got his business acquired while we supported him the whole way through, and he’s still a customer. Forty thousand-plus hours of billable work sits under all of it.

That longevity is the product. Not a burst of output, not a headcount — a decade of someone continuing to choose you, renewed every month on purpose because it’s worth renewing. It’s why we don’t do lock-in contracts. If the only reason you’re still here is that you signed something, I’ve already lost the thing I care about.

Destroy all contractors

So the title stands. Destroy the version of the contractor you learned to distrust — the cash-burner, the short-timer, the person hiding behind a lock-in contract. Replace it with a small senior crew, an on-demand CTO, ruthless scope discipline, and design honest enough to not build you a door you can’t read.

The market is spending an enormous amount of energy trying to out-generate that model. It hasn’t managed to. A lean, senior team pointed at the right problem, evaluating its own work at every gate, is still the highest-leverage way I know to turn finite runway into a business that lasts. I’d rather not argue that in the abstract, though. I’d rather show you the work I already trust.