Faster Than the Ground Can Hold

I had a conversation with a CTO last week that I have been thinking about ever since.

He was describing the pace he is building at, in the private exhaustion of someone who has been moving too fast for too long and has stopped being able to hide it. The public version, with the adoption curves and traction slides, had long since fallen away from the conversation. He was telling me how many hires he has to make by next quarter, which features have to ship before a conference, what competitive pressure looks like now versus six months ago. Every detail was quantified, every deadline already pinned to a roadmap.

Near the end, when the topic had drifted past the immediate pressure, he said something I suspect is being said in more rooms than any of us realize.

“I don’t know if we will even need websites in five years.”

He said it the way someone mentions a weather pattern, with no drama, no thesis, just an observation from a person building a SaaS product, asking whether what he was building would exist in the near future.

Then he kept going. What happens, he wanted to know, when everyone has a personalized agent, and those agents are the ones transacting with other agents on our behalf? What does a landing page do in that world? Or a dashboard? Or SaaS itself, as a category, when the interface layer we have spent fifteen years perfecting turns out to have been the temporary part?

I did not have answers.

He was running faster than anyone I had talked to in months, on a roadmap he had defended in front of a board, toward a product he was privately unsure the world would still need. He is not alone. Every founder I have talked to this quarter, if the conversation goes past forty-five minutes, arrives at some version of the same question.

We are building the bridge, and we are watching the moat crumble.

The velocity is the only part of the job he can still control.

This is the part of being a builder right now that is new, or at least newly sharp. What is different is the rate at which the form factor is moving relative to the build cycle. In a saner era, a founder could reasonably assume that the category they were building in would still resemble itself by the time the product shipped. You were shipping into a known world. The unknowns were demand, distribution, execution; the usual suspects. The ground underneath the product was mostly stable.

The ground is no longer stable. The shape of the interface layer, the role of websites, the fate of search, the future of the buying cycle, the mechanics of how software is discovered and purchased and used; all of it is in open question. The timeline has compressed from decades to weeks.

He is shipping anyway. They are all shipping anyway. The Series A was raised, the team was hired, the pipeline was built, the roadmap was committed to in writing and in public. You cannot easily pause a company because the category it lives in might reorganize during the build.

So they keep scaling, hiring, shipping, and selling, and privately, off the record, they ask each other what the world they are shipping into will actually look like.

I have started to think of this as the specific paradox of the 2026 builder. The pressure to move faster than ever, toward a destination that is moving faster than ever. The velocity is the only legible thing. Everything else requires an admission that nobody is ready to make inside their own company.

I keep thinking about what actually gets built in a condition like this.

When you know the form factor might dissolve, what do you bet on? The surface, or the substrate? The user interface, or the protocol? The brand, or the data model underneath?

Some builders, the ones I have the most quiet respect for, are betting on the substrate. They are building the thing that would still have to exist in five years regardless of the interface. If agents transact with agents, somebody has to be the trusted authority over a particular kind of data, the trusted orchestrator of a particular workflow, the trusted arbiter of a specific decision. The interface can disappear. The trusted authority still needs a home.

Others are betting on the surface. They know the surface will change, but they also know that owning the current surface buys them a position from which to pivot when the next one emerges. Ship the website, win the category, use the momentum to fund the build of whatever comes next.

Both strategies have a logic, and both have been working. The difference is that the people I talk to, the ones running both kinds of companies, now hold their strategies more lightly than they used to. There is a humility in the way they describe their roadmaps that was not there eighteen months ago. They have stopped sounding like they know exactly what they are building toward.

The CTO told me, almost in passing, that his three-year plan now contains a clause about revisiting itself if agent-to-agent commerce becomes real by next quarter. A year ago, that would have sounded like a failure of conviction. Now it sounds like responsible planning.

What I keep coming back to, after that conversation, is the question none of us has quite named out loud.

Who do you become, as a builder, in a season where the only thing you can credibly commit to is the next ninety days?

The identities founders have traditionally built around, the five-year vision, the category definition, the bet on where the world is going, are harder to hold now. The horizon is no longer a place you can confidently describe to your team or investors. You can describe the next sprint, the next quarter, the principles you operate by. Describing the destination has become an exercise in intellectual honesty that not every room rewards.

Some founders are finding a kind of peace in this. They have stopped trying to predict the far horizon and started building things that would be useful to someone in a range of possible futures. They are building for optionality, for defensible primitives, for the few things that do not change even when everything else does.

Others are carrying the weight of still needing to sound certain in rooms where certainty is the currency. They are raising capital on three-year visions they privately do not believe anyone can honestly hold, and defending roadmaps they themselves have started to hedge internally. The gap between the public and private version of their company’s future is widening, the same way the gap between the public and private versions of the AI conversation has widened.

The public narrative still requires conviction; the private reality requires holding uncertainty.

The founder sits between those two, alone, shipping anyway.

I do not have a tidy answer to the CTO’s question. Neither does he. As far as I can tell, neither does anyone in the room where these things actually get decided.

What I am still sitting with is the question of what you build for when what you are building might not exist in its current form in five years. The answer, I am beginning to suspect, is that the real thing being built is the person doing the building. The product may change beyond recognition; the interface might disappear entirely; the category you picked could reorganize around something you cannot yet see. The judgment you develop while shipping at this pace, in this fog, with this level of honesty about what you do not know, stays with you regardless of what gets built.

That is a less satisfying answer than the old versions of founder identity offered, which usually promised that if you picked the right category and shipped the right product, you would end up in a specific place. The new answer is quieter. You might not end up anywhere you currently imagine. You will, however, end up as someone.

The work is shaping the person faster than the person is shaping the work.

I watched the CTO slow down for a minute when that landed in the conversation. He was still going to ship and hire and defend the roadmap in the next board meeting. He just looked, for that one minute, like someone who had been given permission to admit that what he was actually building was himself, and that everything else, the SaaS, the website, the dashboard, the three-year plan, was scaffolding around that.

Which is maybe what it has always been.

Building anyway,

Yen