We Just Resurrected an 18-Year-Old App — And It's Better Than It Ever Could Have Been
In 2008, I submitted Carticipate to the App Store. It was one of the first rideshare apps on the iPhone — maybe one of the first 500 apps, period. The pitch was straightforward: people drive the same routes every day, mostly alone, burning gas and money they don't need to burn. What if they could find each other? Share the ride, split the cost, do something marginally good for the environment along the way.
The idea was right. I still believe that. But building it in 2008 meant asking users to do an enormous amount of manual work just to get started. You had to create a profile. You had to describe your commute — origin, destination, schedule, preferences — in enough detail that a stranger might trust you enough to share a car. You had to check back regularly, search for matches, reach out, coordinate. It was less like using an app and more like managing a small project.
Most people tried it once and moved on. I can't blame them. The friction wasn't a bug in the implementation — it was structural. Matching two people who share a route requires information, and in 2008 the only way to get that information was to ask users to type it in. Some did. Most didn't. Without enough people on both sides, the matches were sparse. Without matches, there was no reason to come back. The loop never closed.
The app quietly died. I moved on to other things, but the idea never fully left. Every time I sat alone in my car on the highway, watching a hundred other solo drivers doing the same thing, I thought about it. The problem was real. The solution was right. Something else had to change.
What changed was AI. Not incrementally — fundamentally. The thing that killed Carticipate in 2008 was the interface: we had to ask users to behave like database operators, filling in fields so the system could do its job. AI eliminates that entirely. You don't fill out a form. You talk to Carla. She asks where you're going, gets the context she needs conversationally, and handles the matching from there. The information that used to require a structured form now comes out naturally in a thirty-second conversation.
The moment it really clicked for me was watching Carla work through a complicated Bay Area commute question — multi-stop, flexible timing, specific neighborhood preferences — in about thirty seconds, the way a knowledgeable local friend would handle it. Casual, specific, useful. In 2008, building that interaction would have taken a developer a month, and it still would have felt robotic. Today it just works.
I'm not sentimental about old software. But some ideas outlast the technology of their moment. Carticipate was one of them. The problem it was trying to solve hasn't changed. The cars are still mostly empty. The commutes are still mostly solo. The waste is still real.
We're back. And this time the technology is ready.