The smell of diesel exhaust is remarkably clingy. It’s heavy, oily, and it sits in the back of your throat like a physical weight. I’m standing on the corner of 4th and Main, watching the taillights of the 42 Express fade into the gray drizzle of a Tuesday afternoon. I missed it by exactly ten seconds. I could see the driver’s profile; I could see the condensation on the glass. He definitely saw me. But the schedule is a god, and the schedule demanded he pull away while my hand was still mid-wave.

There is a specific kind of coldness in being handled with perfect, algorithmic precision. The bus was on time. The system worked. But I am still standing in the rain, and the “system” doesn’t have a mechanism for empathy. It only has a mechanism for throughput.

I see this same coldness, albeit wrapped in much prettier colors, every time I look at the current state of digital health.

The Dopamine Sprout

Take Aisha, for example. Aisha is a composite of about a dozen people I’ve talked to lately, but her experience is singular in its frustration. She’s sitting on her sofa, and her phone vibrates with a cheerful, haptic “thrum.” It’s a notification from her health app. “You’re on a 14-day streak! Keep up the great work, Aisha!” There’s a little animation of a seed sprouting into a flower. It’s beautiful.

It’s delightful. It’s engineered by people who spent six months studying the exact shade of green that triggers a dopamine release in the human prefrontal cortex. Aisha feels a momentary surge of accomplishment. She taps the notification, scrolls through a beautifully rendered dashboard of her “health journey,” and closes the app.

It’s only twenty minutes later, while she’s making tea, that the realization hits her like a splash of that cold Tuesday rain: She hasn’t actually spoken to a human clinician in three months.

Her prescriptions are being refilled by an automated system. Her “concierge” is a chatbot that responds to her concerns about fatigue with a pre-written script about “prioritizing self-care.” The interface is incredibly present. The care is almost entirely absent. This isn’t an accident. It’s an architectural choice.

The Structural Reality

In the world of structural engineering, there’s a guy I know named Thomas B. He’s a bridge inspector. Thomas spends his days hanging off the sides of overpasses, scraping at iron with a specialized hammer. He told me once that the most dangerous bridges aren’t the ones that look scary and rusted; they’re the ones that have been recently painted over without any structural remediation.

“You can make a bridge look brand new with ten gallons of high-gloss epoxy,” Thomas said, wiping grease off a wrench. “But if the rivets underneath are shearing off because of salt corrosion, that paint is just a colorful shroud. It actually makes my job harder because I can’t see the failure until it’s catastrophic.”

– Thomas B., Bridge Inspector

People drive over it feeling safe because it sparkles in the sun. They don’t know they’re riding on a prayer. Most modern health apps are high-gloss epoxy.

Unit Economics vs. Human Nuance

Capital in the tech world flows toward the things that can be measured and scaled instantly. You can measure “Daily Active Users.” You can measure “Retention Rate.” You can measure “Scroll Depth.” You can’t easily measure the nuance of a doctor noticing a slight tremor in a patient’s hand or the way a patient’s voice cracks when they talk about their stress levels.

So, the incentives pool at the surface. The venture capital goes to the UI/UX designers who can keep Aisha clicking. It doesn’t go to the unscalable, expensive, and deeply human work of clinical oversight. The app is addictive by design because your subscription fee is the product. Your health outcome is just a secondary, often ignored, byproduct.

To understand how this actually works, you have to look at the “Unit Economics” of a telehealth visit. In a traditional clinic, a doctor’s time is the primary cost. In a “disruptive” health tech startup, the doctor’s time is a “cost of goods sold” (COGS) that needs to be minimized to achieve profitability. Every minute a licensed physician spends talking to you is a minute that eats into the company’s margins.

The Feedback Loop of Ignorance

The goal of the software, therefore, is to act as a filter-to keep you away from the doctor for as long as possible. They call it “automated triaging” or “asynchronous care,” but often it’s just a sophisticated way of ghosting the patient. The “streak” and the “gamification” are the distractions. They are the shiny red paint on Thomas B.’s crumbling bridge.

I’ve made the mistake of trusting the gloss before. I once used a fitness app that congratulated me on my “consistency” for three weeks straight while I was nursing a stress fracture that I was too stubborn to admit was there. The app didn’t know I was limping; it only knew my GPS was moving. It rewarded the movement and ignored the injury. It was a perfect feedback loop of total ignorance.

The frustration Aisha feels is the “clinical gap.” It’s the space between the promise of the marketing and the reality of the treatment. When you realize that the company cares more about your “engagement” than your “endocrinology,” the sprout animation starts to look a lot more like an insult.

A Different Foundation

There is a different way to build these systems, but it requires a fundamental rejection of the “app-first” mentality. It requires putting the physician back at the center of the wheel, rather than treating them as a backend API that the software occasionally calls upon.

This is the model that Mochi Health was built upon. Instead of a revolving door of disconnected providers and automated “check-ins” that lead nowhere, the focus shifts to continuity.

Continuity of care is a boring term in the tech world. It doesn’t sound “disruptive.” It sounds like something from 1955. But in medicine, continuity is the only thing that actually works. It’s the difference between a doctor who knows your history, your lab results, and your specific hurdles, and a chatbot that suggests you “drink more water.”

When you have a coordinated team-doctors, nutritionists, and real support staff-the app becomes a tool for communication rather than a substitute for it. I think about the bridge again. If Thomas B. were designing a health app, he wouldn’t start with the animations. He’d start with the rivets.

Quality is Messy

We have been trained to equate “ease of use” with “quality of care.” If the app is smooth, we assume the medicine is sound. If the reorder process is one-click, we assume the oversight is thorough. But these are two entirely different skill sets. Building a great e-commerce interface is a solved problem. Providing comprehensive, physician-led obesity management or primary care is an incredibly difficult, messy, human problem.

The danger of the “delightful interface” is that it creates a false sense of security. It’s a “deferred tax” on our health. We pay for the convenience now with the lack of outcomes later. We feel cared for because the app remembered our birthday, while our actual metabolic health remains a mystery because no one has looked at our bloodwork in six months.

The Wrong Metric

I’m still thinking about that bus. The driver wasn’t a bad person; he was just optimized for a different metric than “passenger satisfaction.” He was optimized for “schedule adherence.”

When you sign up for a health service, you have to ask yourself: What is this system optimized for? Is it optimized to keep me well, or is it optimized to keep me subscribed? If the app feels like a game, you might be the one being played.

The most revolutionary thing a digital health company can do right now isn’t to add more AI or more “streaks.” It’s to admit that the software is the least important part of the equation. The software should be invisible.

It should be the quiet, efficient conduit that allows a real doctor to do real work. It should be the bridge that actually carries you across, not just the paint that makes you think you’re moving.

Conclusion: Choose the Rattling Bus

I finally caught the next bus. It was twenty minutes late, and the heater was broken, and the seat was slightly damp. It was a sub-optimal experience by every digital metric. But it got me where I needed to go. It was a physical, clunky, imperfect solution to a real-world problem.

In health, we should be wary of anything that feels too smooth. Real healing is often clunky. It’s a series of difficult conversations, adjusted dosages, and labs that don’t always say what we want them to say. It doesn’t always fit into a 7-day streak.

But at the end of the day, I’d rather be on a rattling bus heading toward my destination than standing on a beautiful, high-tech sidewalk watching the future pull away without me.