First, a quick apology for the unexpected hiatus. Life has a way of sneaking up from behind with surprises you can’t always brace for.
Now, for today’s Napkin Sketch, I want to start with a short story.
Last weekend, my wife and I were driving back to the Bay Area from a quick Orange County getaway. I was borrowing my mom’s Mercedes. Near the tail end of the trip, a little coffee cup icon popped up on the dashboard, telling me to take a break.
Curious, I asked my wife to Google it. Turns out, this wasn’t just a timer-based “you’ve been driving for X hours” kind of thing. It was Mercedes’ Attention Assist, a system that, in the first few minutes of your drive, builds a driver profile from over 70 parameters — things like steering micro-corrections, lane positioning, pedal pressure, and more. Then it compares your real-time driving to that baseline. When it sees enough deviation, it knows you might be getting drowsy… and that coffee icon lights up. Love it.
Credit where it’s due, Mercedes has nailed it with this feature. It’s an impressive showcase of ambient cognition in action — technology passively collecting signals in the background, interpreting them, and surfacing a timely intervention without you ever needing to ask.
There’s just one issue: I didn’t feel tired. At all. And as far as I could tell, my driving was perfectly fine.
That’s when the thought hit me: this is the exact tension we face when building neurotech products — the friction between what the system detects and what the human believes.
We have two competing realities here:
The machine’s perspective: built from dozens of micro-movements, reaction timings, and subtle deviations, all invisible to the human eye (or ego).
The driver’s perspective: “I feel fine, I know myself, and I don’t need some coffee cup telling me what to do.”
In driving, this tension is an inconvenience at worst. You can ignore the icon and keep going. But in neurotech, where the signals are about your mental state, this friction gets even trickier.
Imagine a wearable or app that builds your “cognitive baseline” — the brainwave equivalent of that driver profile. It notices that your focus is slipping, your reaction times are slowing, your typing cadence is changing… and it tells you, “Take a break.”
Now imagine you don’t feel that way. You’re in the zone, cranking through emails or deep in design work. Do you trust the tool? Or do you dismiss it, thinking, “Not now, I’m fine”?
The irony is, the moments when you think you’re fine are often the exact moments you’re not. Mental fatigue can creep in so subtly that you only notice it after a mistake, a forgotten detail, or a sudden dip in productivity. That’s the entire promise of ambient cognition systems: to catch what you can’t.
But here’s the hard part:If the feedback is too often wrong (or feels wrong), the tool loses credibility. The magic of a well-timed nudge becomes the annoyance of an unwanted interruption. Trust erodes fast.
The sweet spot — and the real product challenge — lies in making those nudges feel uncannily accurate, like the system is reading your mind (or your mind’s micro-signals) better than you can. When it’s right, it feels like a superpower. When it’s wrong, it feels like spam.
So, as we design the next generation of cognitive feedback tools, here’s the challenge:
Build baselines that are sensitive enough to catch subtle shifts without overreacting.
Deliver feedback in ways that feel helpful, not judgmental.
Give people context so they understand why the nudge appeared — and maybe even start to notice the signals themselves.
The moment someone goes from “Why is it telling me this?” to “Ah, I see what it caught there” is the moment they start trusting it. And once trust is established, that’s when the real value kicks in.
Mercedes built a coffee cup for your driving state. Neurotech will build one for your mind.
The question is: will you believe it when it lights up?