You’ve probably noticed it by now. The late-night ping to your AI assistant instead of the Slack message to your colleague. The immediate, frictionless answer that saves you from a two-hour meeting. It feels like a superpower.
But it’s actually a trap.
Convenience is the most dangerous drug in knowledge work because it feels exactly like progress.
It’s called the Waymo effect. When a self-driving car removes the friction of dealing with a cab driver, you get to your destination in silence. When an LLM removes the friction of dealing with a colleague, you get your research in silence. You don’t have to explain your premise. You don’t have to negotiate methodology. You don’t have to align with another human being’s perspective. The AI just does it.
Most critiques of AI focus on accuracy, bias, or job loss. But the deeper risk is structural. AI optimizes for the individual’s task and externalizes the cost of collaboration. The less social the research process becomes, the more productive it looks by conventional metrics.
When you remove the friction of dealing with other humans, you don’t just remove the annoyance—you remove the serendipity.
Think about the last time you hit a wall on a project. Did you walk down the hall to argue with a peer, or did you prompt a model? The AI agreed with your premise, filled the gaps, and let you move forward. You were highly efficient. But you were also completely alone.
The messy arguments in the lab, the debates over coffee, the forced negotiations over methodology—those weren’t inefficiencies. They were the forge where trust, serendipity, and true intellectual convergence were built. By outsourcing our friction to machines, we are starving the very human interactions that make research credible and worth building on.
The smoother your research process feels, the more it is quietly starving the human friction that produces true insight.
People are becoming islands, working on isolated tasks without the bigger picture. We are optimizing for the individual task and destroying the collective intelligence.
Convenience is not neutral. Adopting AI to escape human friction trades away the collaborative context that makes knowledge work adaptive and real. The next time you open a prompt box instead of walking to a colleague’s desk, remember the trade-off.
If your AI assistant never disagrees with you, you aren’t doing research. You’re just writing a diary that code can read.
FAQ
Q: Isn't AI just removing unnecessary middlemen and bureaucratic friction?
A: No. While it removes annoying meetings, it also removes forced alignment. The 'unnecessary' arguments are often where hidden assumptions are exposed and true breakthroughs happen.
Q: How do I use AI without becoming an isolated island?
A: Use AI for execution, not for ideation or peer review. Keep the messy, human arguments in the loop. If a task requires negotiation or consensus, don't outsource it to a machine.
Q: Isn't most human collaboration in research actually just ego and busywork?
A: Sometimes. But that ego and busywork is the price of trust. Without it, you get a collection of highly efficient individuals working in silos, producing work no one else can safely build upon.