Why Data Products Break the Binary

SLOW DOWN TO GO FAST

There’s an old engineering adage: Slow down to go fast. Study the problem before you solve it. Understand the system before you optimize it. Create meaning before you scale.

Everything about hustle culture militates against this. The pressure to ship, to grow, to demonstrate traction for the next funding round creates a peculiar form of gaslighting in which founders feel inadequate for doing the actual work of understanding. The hustle becomes the point, disconnected from any coherent theory of value creation.

But here’s what I’ve noticed: When founders actually disconnect from the main artery, when they stop refreshing metrics dashboards and start talking to users, stop chasing vanity milestones and start studying interaction patterns, they report a sense of clarity. The fog lifts. The cognitive dissonance resolves. This isn’t mysticism. It’s what happens when you stop trying to force a complex system into a complicated framework.

The anxiety of chasing an unreachable binary state gives way to the quieter work of tending emergence. Data products reward this patience. Because the data changes, because human interaction with it evolves, the founders who take time to understand what’s actually happening, not what the PMF playbook says should happen, are the ones who build systems that endure.

THE COGNITIVE DISSONANCE OF DATA FOUNDERS

Here’s what I witness constantly: founders with working products and paying customers who cannot proceed because they lack seed funding. They’ve built something people want and will pay for. Yet the mythology tells them they’re failing because they haven’t secured institutional validation from investors optimizing for hockey sticks and blitzscale.

VCs want exponential “hockey stick” growth that comes from funding money, not a tried-and-true but plodding PMF playbook. Seed funding is also tied to the mythos that founders who have a working product and revenue are not doing it right: Where is their “Make me rich overnight campaign” I can get all my investors behind? Why do you need seed funding to build a product if you already have one?

The VC model wasn’t designed for data products. It was designed for software that ships and scales, code that does the same thing for customer one and customer 1 million. But a data product that does the same thing for every customer is, almost by definition, failing. The value proposition is a contextual response, not a uniform delivery.

This creates devastating cognitive dissonance. Founders who understand their users, who iterate based on actual interaction patterns, and who resist the temptation to scale prematurely are often perceived as “slow” by PMF standards.

They haven’t “crossed.” They’re still in the chasm. Except there is no chasm. There’s a river, and they’re learning to navigate it.

HOW WE PROCEED AT BAST

We don’t pretend to know what PMF looks like for an AI system that helps clinicians support traumatic brain injury recovery or analysts evaluate sports performance. The honest answer is that we can’t know in advance, because the data will change, and human interaction with that data will change both humans and the system.

Instead, we create conditions for emergence:

Define a target audience. Not the 80% who might love your product someday, but specific humans with specific needs encountering particular information.

Identify authoritative sources. What does this audience actually need access to? What knowledge, curated from what origins, with what provenance, serves their real work?

Build transparent systems. Every recommendation is traceable to its sources. Every confidence score is grounded in actual authority. Every extrapolation beyond the training data is explicitly flagged.

Then experiment. Watch what happens. Use the interaction data to define the subsequent use case and to understand how the conversational AI should evolve as the dataset grows.

The system learns; the humans learn; the data evolves; the cycle continues. This isn’t crossing a chasm. It’s tending a garden in shifting soil.

THE MAINSTREAM FUTURE

The organizations that thrive with data products won’t be those that achieve PMF as a binary state. They will be those who develop the organizational capacity to work with uncertainty, to probe, sense, and respond.

In my mental health column, I argued that the future belongs to transparent systems that cite authoritative evidence, that show their work, that help humans become active participants in their own understanding rather than passive consumers of algorithmic output.

The same principle applies across all domains in which data is the primary input. Explainable AI isn’t a nice-to-have feature. It’s the infrastructure that enables continuous co-evolution.

When humans can see why a system made a recommendation, they can contextualize, challenge, and refine it.

The interaction becomes generative rather than extractive. Whittaker told the World Economic Forum: “We can’t build new worlds, we can’t imagine new paradigms without that incubation space, without the safety to experiment with ideas and think about what could work or not.”

That safety doesn’t come from achieving fit. It stems from building systems that are transparent enough for humans and data to evolve together.

The chasm was always a metaphor for a product world. Data products need a different metaphor—not a gap to be crossed, but a current to be navigated. Not a binary arrival, but a continuous relationship.

The organizations that understand this, that stop chasing the binary and start creating conditions for emergence, will be the ones still standing when the trillion-dollar infrastructure bets collapse under their own contradictions.