The Paradox Every Founder Hears

"Talk to customers," they said.

"Find what people want," they said.

"Distribute before building," they said.

Then came the counter-canon: Ford's oft-repeated quip that customers just wanted "faster horses," and Steve Jobs's famous line that people don't know what they want until you show them.

So which is it?

The advice feels schizophrenic. One camp preaches obsessive listening, the other demands blind conviction. If you're building, which god do you worship? The one who says ask the market, or the one who says ignore it?

The truth, as every seasoned founder eventually learns, is that this isn't a choice at all. It's a sequence. You listen to reveal the job. You lead to deliver the solution. And the real art, the difference between vitamins that collect dust and painkillers that customers reorder, is knowing how to toggle between the two.

A Founder's Realization

I remember a moment in my own journey where this tension crystallized. I had built a feature we assumed was a "nice to have." A customer shrugged at it during a demo. But a month later, they were still using it daily and building processes around it.

That was the shift. They hadn't asked for it. They didn't praise it. But retention revealed the truth: for them, this was no vitamin. It was morphine.

That realization is the emotional pivot every founder goes through. It's when you stop taking customer words at face value and start studying customer behavior. Customers articulate symptoms, not cures. They ask for horses; the real job is to get across town faster.

The Frameworks That Clarify the Fog

If you zoom out from that founder's moment, several canonical lenses snap into place.

• Product/Market Fit (Marc Andreessen): markets set the slope; products set the intercept. The "listen vs. lead" debate sits entirely within this constraint.
• Jobs To Be Done (Clay Christensen): people "hire" products to make progress in specific circumstances. They don't want features, they want outcomes.
• Retention as the Tell (Alex Schultz): acquisition can be bought; retention is earned. If cohorts flatten and expand, you've found pain.
• Superhuman's PMF Engine (Rahul Vohra): measure "very disappointed" users, segment ruthlessly, and build for intensity of love.

Each of these frameworks reframes the paradox: listen to customers not for the solution, but for the job; lead them not with arbitrary vision, but with a sharper answer to the job than they could imagine.

Vitamins, Painkillers, and the Escalator

The vitamin vs. painkiller metaphor survives because it's directionally right. Vitamins are aspirational, optional, easy to skip. Painkillers are urgent, irrefutable, hard to live without.

But the real trick is designing the escalator between them:

1. Tight JTBD fit (address the real job).
2. Faster time-to-value (customers feel relief quickly).
3. Repeatable ROI (outcomes are provable, not anecdotal).
4. Embedded workflow (switching costs grow).
5. Compounding distribution (network effects, integrations, loops).

This escalator explains why some "toys" turn into billion-dollar tools, while others fade as demos.

Case Studies: How Today's Winners Made the Turn

Cursor (Anysphere)

Dismissed early as "just autocomplete," Cursor transformed when it collapsed the entire coding loop editing, testing, debugging into one AI-first environment. Once developers stopped switching back and forth, the tool went from vitamin to painkiller. A $9B valuation later, the verdict is clear: necessity, not novelty.

Bolt.new (StackBlitz)

Initially a curiosity ("AI builds an app from a prompt"), Bolt looked toy-like. But when teams began shipping production apps in days instead of weeks, the job clarified: collapse scaffolding time. ARR ramped in weeks. A toy became a must-have engine.

Lovable

"Chat to build apps" sounded like a demo until it met a hair-on-fire job: ship a working v1 without a dev team. The job wasn't "play with AI," it was "launch fast or die." Once founders anchored on that urgency, Lovable ratcheted from novelty into necessity.

Firecrawl

At first, an esoteric scraper. But when LLM agents scaled, a bottleneck appeared: messy, unreliable web data. Firecrawl's clean pipeline (scrape → markdown/JSON) became the invisible painkiller. Nobody asked for it, but once adopted, nobody built without it.

Airbnb & Uber (contextual anchors)

Both began as fragile "toys." Couch-surfing with strangers? Hailing cars from your phone? Skepticism was the norm. The shift came when they ratcheted the job: lodging when hotels are full; transport you can summon reliably. From there, they locked in distribution and never looked back.

The Lighthouse Checklist: Five Tenets

From these stories and frameworks, a field manual emerges:

1. Anchor on a job, not a feature. Phrase it in a CFO's language: "reduce time-to-X by Y%."
2. Segment ruthlessly. Chase intensity, not averages. 100 fanatics beat 10,000 tourists.
3. Design the vitamin → painkiller escalator. Make the transition explicit in roadmap and metrics.
4. Prove retention before scaling acquisition. Paid growth is gasoline; retention is the engine.
5. Price outcomes, not features. Charge for the pain you remove, not the bells you add.

This isn't "best practices." It's the common DNA across winners.

The Antifragile Reveal

Here's the deeper truth: the founders who thrive aren't the ones who guess right once. They're the ones who structure their systems to benefit from randomness.

Nassim Taleb calls this antifragility: systems that grow stronger from volatility.

• Tinkering > Planning: Firecrawl didn't plan to be indispensable. It tinkered until the job revealed itself.
• Optionality > Certainty: Cursor, Bolt, Lovable looked like side projects until timing turned them. Multiple small bets kept them alive long enough to matter.
• Ratcheting > Recycling: Once Airbnb proved network effects, they ratcheted upward into a hospitality empire. No backsliding to hobbyist couch-surfing.

That's the meta-lesson. Don't worship Ford or Jobs. Don't obsess over vitamins vs. painkillers. Build antifragile companies: ones that experiment cheaply, preserve optionality, and ratchet when they hit gold.

Let's close the Loop

So, which was right Ford's horse line or Jobs's provocation? Both, and neither. Customers describe symptoms; founders must interpret jobs. Customers can't imagine solutions; founders must lead them there.

But the founders who endure are the ones who stop treating this as a binary choice and start treating it as a system design problem. They tinker, preserve options, and ratchet successes.

That's how vitamins become painkillers. That's how "toys" become inevitabilities. That's how businesses, like systems in Taleb's world, become antifragile.

And that's the part they never tell you on the podcast.

Further reading

• Final Journey of Job Titles to the age of Sovereign Work
• Do Fulfilling Work
• Uberization of Knowledge Work