In 2010 I watched three strangers at a bar scroll instead of talk, and knew social media had crossed from early adopters to everyone. I had the same n=3 moment about AI this week — and the signal that matters isn't the founder shipping an agent. It's the former physician in Ecuador using ChatGPT every day.
Circa 2010, I was sitting at a bar with my cousin watching three twenty-somethings at the other end of it. They never said a word to each other. They scrolled their phones and occasionally turned a screen around to show off something they'd found.
That was an n=3 sign to me that something major was changing.
2010 was the year social media really blew up. Facebook apps were the rage. A high Klout score was your ticket into certain jobs. "Social business" was the phrase every agency and consultancy was running with. Everyone was photographing their lunch for Twitter.
Three strangers not talking at a bar is not data. It's an anecdote. But it was the anecdote that told me the behavior had left the early-adopter cul-de-sac and entered the bloodstream. The numbers caught up later. The bar got there first.
I had the same moment this week
I was out to lunch with my wife, my brother-in-law, and my sister-in-law. Half the conversation was about AI. Same thing all week. Every time my wife and I go out for a walk or a coffee, AI comes up.
My brother-in-law mentioned that someone at his dental office — a former physician in Ecuador who had never touched ChatGPT — now uses it constantly. For work, for play, for everything. My wife and sister-in-law are working out how to make parenting easier with Claude. I'm the odd one out: I barely open social media anymore, and most of my time outside work is spent talking to AI.
n=3 again. Same sample size. Same feeling.
A former physician in Ecuador using ChatGPT every day is a different signal than a YC founder shipping an agent.
The signal isn't the builder
Most of the AI adoption discourse measures the wrong room. It tracks enterprise rollouts, capability benchmarks, and what the builder community is shipping. Those are real, but they're early-adopter telemetry. They tell you what the people closest to the technology are doing.
The 2010 inflection wasn't early adopters spending more time online. It was average consumers starting to. The tell was never the Klout obsessives. It was my aunt joining Facebook to see photos of her grandkids.
The AI version of that tell is the former physician in Ecuador. It's the parent reaching for Claude instead of a forum. These people have no reason to care about model architecture or agent frameworks. They're not adopting AI because it's AI. They're adopting it because it's the fastest way to get something done. That's what crossing the chasm actually looks like from the ground.
Pattern recognition is a leadership skill, not a vibe
It's tempting to file n=3 moments under intuition and move on. But knowing which weak signals to trust is one of the most underrated parts of leading product. The data is always late. By the time consumer AI usage shows up cleanly in a chart, the window to act on the insight has mostly closed.
The discipline isn't trusting every anecdote. It's the opposite. Most n=3 moments are noise — three people who happen to share your bubble. The 2010 bar worked as a signal precisely because those three strangers were outside my world. This week's lunch works because the former physician is nowhere near tech.
So the question I keep asking, of myself and the PMs I work with: when you get a strong gut read from a tiny sample, is the sample inside your bubble or outside it? Inside, discount it. Outside, pay attention. The signal that matters is almost always the one coming from someone who has no reason to be an early adopter — and is one anyway.
2010 redux.
The behavior has crossed over. Not the builders. The everyone-else. If you lead a product right now and you're still calibrating to what the AI-native crowd is doing, you're reading last year's room.
Find the people in your life with no reason to care about this technology. Watch what they reach for. That's the inflection, and it's already here.
Frequently Asked Questions
What is an n=3 signal?
An n=3 signal is a tiny, unscientific sample — three people, one room — that nonetheless tells you a behavior has crossed from early adopters into the mainstream. It isn't proof. It's the moment your intuition catches a shift the data won't confirm for another year. The skill is knowing which n=3 moments to trust.
Why does consumer AI adoption look like 2010 social media?
In 2010, average consumers — not early adopters, not businesses — started spending real daily time on social platforms. Consumer AI is hitting the same threshold: non-technical people (parents, healthcare workers, service workers) are now using tools like ChatGPT and Claude every day for ordinary tasks. The tell isn't capability benchmarks; it's everyday usage by people with no reason to care about the technology itself.