The Friction You Are Removing Was Doing Something
TLDR card reading: Great teams live in the spaces between people. Solo AI use optimizes those spaces away.
At Nivan Live 2026, Casey Hudetz opened his talk by repeating a sentence he had been hearing everywhere.
"Now I don't need to bug someone."
He said it three times. Once about brainstorming, because a large language model will hand you forty ideas before a meeting invite gets accepted. Once about research, because you can ask a tool for a synthesis instead of asking the person who ran the study. Once about design, because you can generate a screen for a pitch deck without waiting on anyone.
Then he named what that adds up to. We are becoming, he said, a bug free workforce. We do not have to bug each other as much anymore.
He is right that this is a gain. Our toolkits are bigger and our roles have expanded. But I have not stopped thinking about the question he asked next, which is what happens to a team when the reasons to interrupt each other quietly disappear.
What the friction was carrying
Casey reached for Sherry Turkle, who has spent two decades arguing that we expect more from technology and less from each other. Then he reached for evidence.
Google's Project Aristotle studied 180 teams to work out what separates the high performing ones. The answer was not talent density or resources. It was psychological safety, the sense that you can take a chance, push back, or fail without it costing you. That kind of safety is not declared in an offsite. It accumulates through small, low stakes interactions that mostly look like nothing.
Then there is the study I keep returning to. Researchers at MIT, working with a Bank of America call centre, fitted employees with sociometric badges to map who actually talked to whom. Breaks had been staggered for efficiency, so coverage never dipped. The researchers suggested the opposite: let each team break at the same time.
Productivity rose by roughly 15 to 20 percent. The lowest performing teams cut their average call handling time by more than 20 percent. Stress fell. The bank estimated the effect across its call centres at around 15 million dollars a year.
Nobody was trained. Nobody got a new system. They were simply allowed to have coffee together.
What was moving in those breaks was tacit knowledge. How one person calmed an angry caller. A faster route through a slow screen. The unwritten interpretation of an unclear rule. None of it was in the handbook, and none of it would have survived being written down.
The part that is easy to miss
The friction AI removes is not always waste.
Some of it was the interaction where trust got built. Some of it was where context got shared. Some of it was where a junior person learned by watching someone senior think out loud, which is a form of teaching that has never had a name and never appeared on a roadmap.
Casey made a point about mentorship that I found harder to shake than the statistics. When he turns to a model with a career question instead of a person, he is not only losing the answer that person would have given. He is taking away their chance to crystallise what they know and pass it on. The mentor loses something too.
This is the version of AI risk that does not make headlines. Nothing breaks. Nobody complains. The team simply becomes slightly less of a team, one removed interruption at a time, and by the time it shows up in retention numbers the cause is untraceable.
What to do about it
Casey's answer was not to use less AI. It was to be deliberate about which friction you remove.
Find the parts of your work that are genuinely toil, highly manual and low judgment, and delegate those. Then look hard at what is left, and ask whether any of it was quietly doing a second job.
His most useful reframe was borrowed from Pixar, whose original studio Steve Jobs helped lay out so that people from different departments would keep running into each other. Brad Bird put it plainly: when people run into each other, when they make eye contact, things happen. Casey called this productive friction. Not friction you tolerate. Friction you protect.
I would put it this way. As AI removes more of the reasons we have to bug each other, the reasons we choose to become more important. That is a design decision, and somebody has to make it on purpose.
So here is the question I took home from his session, which is the one he closed on.
The next time you turn to a model for an answer, is there someone you could reach out to and bug instead?
Watch the session: Casey Hudetz, Senior Product Design Manager at Docusign, on The Bug Free Workforce. Recorded 21 May 2026 at the Institute of Design at Illinois Tech, Chicago. [Add link]
Sources
Sherry Turkle, Alone Together: Why We Expect More from Technology and Less from Each Other, Basic Books, 2011.
https://www.sherryturkle.com/alone-together
Charles Duhigg, "What Google Learned From Its Quest to Build the Perfect Team," The New York Times Magazine, 25 February 2016. On Project Aristotle and psychological safety.
https://www.nytimes.com/2016/02/28/magazine/what-google-learned-from-its-quest-to-build-the-perfect-team.html
Alex "Sandy" Pentland, "The New Science of Building Great Teams," Harvard Business Review, April 2012. On the Bank of America call centre and aligned coffee breaks.
https://hbr.org/2012/04/the-new-science-of-building-great-teams
Benjamin Waber, Daniel Olguin Olguin, Taemie Kim and Alex Pentland, "Productivity Through Coffee Breaks: Changing Social Networks by Changing Break Structure," MIT Human Dynamics Laboratory, 2010. The underlying paper.
https://www.ssrn.com/abstract=1586375
Brad Bird on the Pixar atrium, quoted in Walter Isaacson, Steve Jobs, Simon & Schuster, 2011, from a 2008 conversation.
Watch the session: Casey Hudetz, Senior Product Design Manager at Docusign, on The Bug Free Workforce. Recorded 21 May 2026 at the Institute of Design at Illinois Tech, Chicago.
