Managing your innovation clock speed

Elon Musk went into the Department of Government Efficiency with the energy – and subtlety – that had built Tesla and SpaceX. A hundred and thirty days later, roughly one product cycle, he was scaling back his time in Washington, saying the effort had made progress but was "not as effective as I'd like". Elizabeth Holmes promised lab results from a single drop of blood and ran into an FDA approval process built for an earlier and more careful age. Peter Thiel's Palantir, meanwhile, has spent two decades winning contracts worth hundreds of millions from the Department of Defense, working with the FDA to speed up its reviews, and taking the largest contract the NHS has ever awarded, all inside some of the most bureaucratic institutions on earth.

Three ambitious innovators, and two very different endings. It is tempting to explain the difference by talent or resources, but that doesn't hold up. What separates them is an understanding of how fast the world around them was able to move, and what to do at the edges where fast meets slow.

Civilization runs at six speeds.

Stewart Brand has a useful way of seeing this, which he calls pace layering. Think of civilization as six layers stacked on top of one another, each moving at its own speed:

• Fashion changes with the season
• Commerce adapts every year
• Infrastructure evolves over decades
• Governance shifts over generations
• Culture transforms over centuries
• Nature moves over millennia

The interesting things, good and bad, happen where the layers meet. If your innovation lives in the commerce layer, with quarterly results and rapid iteration, and what you are trying to change lives in the governance layer, with its approvals and its need for consensus, you are not just meeting resistance. You are pushing against something that moves at a different speed by design.

That was DOGE's problem. Moving fast and breaking things works well when you are building rockets, where you control most of the variables, or taking on the car industry, where market forces eventually override regulatory lag. Government efficiency, however, sits across several layers at once, from digital transformation at one end to constitutional law at the other, and no amount of energy makes constitutional law move at the pace of a software release, which is, on reflection, the point of constitutional law.

Speed is not the same as recklessness.

Theranos is the harder case, because the vision was genuinely good, and a good vision is a dangerous thing to have. Holmes and Sunny Balwani believed they could compress decades of medical device validation into a startup timeline, and they made three mistakes at the layer boundaries that anyone innovating in a regulated field should study.

The first was treating the FDA's pace as an obstacle rather than a signal. Medical devices sit in the governance layer for a reason; the slowness encodes hard lessons about what happens when untested technology reaches patients. Rather than build a bridge across that gap, they went around it, through secrecy.

The second was underestimating culture. Trust in a medical test is built over generations of verified results. Theranos bet that a superior technology could replace that trust in a few years, and culture does not work that way.

The third was dismissing expertise. Clinical laboratories hold deep institutional knowledge because the stakes are life and death. Treated as a partner, that knowledge could have made the vision viable. Treated as a competitor, it made the company an adversary of everyone who could have helped.

They did not fail because they moved fast; plenty of people move fast in medicine. They failed because they crossed layers without respecting how fast each one could go.

How to move fast in a slow world.

The innovators who last do not fight the pace layers. They work with them, and there are three habits worth copying.

Map the layers before you build. SpaceX succeeded partly because Musk understood he would have to work within existing aerospace regulation while showing the FAA that the regulation could evolve. He gave the regulator reasons to adapt rather than reasons to resist. Before you start, ask which layer runs at your natural speed, which layers you have to influence to succeed, and where the friction between them will be.

Build translation interfaces. The most dangerous boundaries are the ones where two speeds meet with nothing in between; think of a shock absorber that lets energy pass without a collision. Netflix did this when it took on television. Instead of attacking the cable companies, it built relationships with content creators at commerce speed, shifted viewing habits at culture speed, and stayed inside telecommunications rules at governance speed. Each layer adapted at its own pace, and the whole thing still moved. In practice this means finding the people who already span the boundary, agreeing on measures both sides recognize, and running pilots that prove the idea without threatening the core system.

Sequence your speed. This is not the same as slowing down. Amazon Web Services was built at commerce speed, with rapid iteration and constant customer feedback; sold to enterprises at infrastructure speed, on multi-year contracts and gradual migration; and negotiated with governments at governance speed, on data sovereignty. Three speeds, one transformation.

What Palantir did differently.

Palantir is the clearest example I know of fast innovation inside slow systems, and its playbook is less mysterious than its reputation, which the company has done little to discourage.

It became native to the boundary. Where Musk mounted a frontal assault and Holmes went around, Palantir embedded itself in how government already works. It did not try to change procurement; it mastered it. Its FDA contract, which exists to make regulatory review faster, is the same idea seen from the inside.

It solved real problems slowly. Thiel has called this patient capitalism: solutions built for specific government needs over years, not quarters. The defense contracts were not won with demos but with the unglamorous work of proving the systems could handle classified data, awkward integrations and bureaucratic oversight.

It built its own translators. Former officials who understood both worlds. Software that fitted into existing government systems instead of replacing them. Development methods that could live inside government approval cycles.

And it played the long game. Each project became the bridge to the next agency, then the next country, across administrations of both colors. Palantir is now among the most valuable companies in the world, and its government work has grown regardless of who is in office.

DOGE is the counter-example, and a recent one. After a hundred and thirty days Musk had walked back the two-trillion-dollar savings target, and former staff expected the effort to fizzle without him. The error was the classic one: commerce-speed methods, meaning public pressure, rapid iteration and bold promises, applied to a governance-speed problem. That approach works when you control the variables. It breaks down in a system built for deliberate, consensus-based change. Government can be improved, of course. But improving it takes a different playbook, one Palantir has spent twenty years writing.

The question is not whether you can move fast.

Whether you are leading a digital transformation in a traditional industry, starting a company in a regulated market, or simply trying to change something inside a large organization, the same four things apply. Map the layers you are trying to influence and the speeds they naturally run at. Design the interfaces that let fast and slow exchange energy without a crash. Move fast where you can, build bridges where you must, and be patient where the stakes are highest. And learn to tell friction that makes the idea stronger from delay that adds nothing.

None of this slows innovation down. What it does is speed up the whole system's ability to absorb it. Holmes tried to bypass the system and became a cautionary tale. Musk tried to shock it and stepped back inside five months. Palantir chose to evolve with it, and built one of the most valuable companies of the decade. So the question is less whether you can move fast in a slow world, and more whether you can help the world move faster without breaking what matters, and whether you have the patience to play that long game.

This is not a commentary on the people or initiatives involved, but on the choices worth considering when you innovate in a context you do not know well.