Hey there,
In 2002, my boss at Sony handed me a task that had no business landing on my desk. Define our approach to platform device management, with the first proof point being an e-commerce interface running on a television set-top box. This was 2002. Nobody had done this. Then he asked me to pitch our group's entire ePlatform initiative, a customer management platform, a web services architecture, and this device management piece among them, to Sony Pictures and Sony Online Entertainment, trying to get two divisions with their own agendas to buy into an approach built by a product guy who, in their eyes, may not have belonged in that room at all.
I had no script for any of it. Nobody handed me a framework. I had to figure out what I actually thought, defend it in front of people much more experienced and much higher up in the organization, each with their own strategic directives, and live with what I got wrong. That discomfort was the point. My boss could see the whole task and the whole person he was handing it to, and he calibrated the difficulty on purpose, the same way it happened later at Qualcomm when a role meant to cover software and user experience for a personal television device quietly grew into owning the hardware and the service as one system, and again at ProQuest when another team landed under me not long after I started.
None of those bosses were being generous. They were betting that I'd rise to something I hadn't done before, and the only way to find out was to make me sit in the not knowing.
A year ago, right around this same time, I wrote The Elul Principle, an argument that AI transformation asks something harder than adding a new skill. It asks you to release the identity that made you successful before AI existed. I wrote about senior people who resist the tools not because they're technically behind, but because the tools threaten a version of themselves built on being the fastest processor of information in the room. I told you what to examine, what to let go of, what to unbecome.
I realize now that I only tackled half of the issue. And so, this week, actually on Yom Kippur itself (I can't pretend the timing is a coincidence), I think I finally see the half I left out.
Clay Shirky, who spends his days as a vice provost at NYU watching what AI actually does inside classrooms, wrote a piece I enjoyed, and it got me thinking. A professor friend of his once learned that her students were typing her own questions into ChatGPT and reading the answers back to her as if they were their own thoughts. Shirky first read that as ordinary cognitive offloading, kids skipping the thinking. Looking back, he's decided it was something else. He calls it emotional offloading, the use of AI to avoid the discomfort of an unscripted human moment rather than the effort of a hard one. Students rehearsing an awkward conversation with AI before having it with a friend. Students who'd rather draft an apology with a chatbot than sit in the discomfort of an actual professor's disappointment. He borrows a term from sociology for what happens next: social deskilling, the erosion of a person's tolerance for the friction that used to be unavoidable and is now, for the first time, optional.
That word optional is the half of my own argument I missed. Elul was about the leader who avoids AI because it threatens who they are. What I didn't ask is what that leader loses the ability to see because they avoid AI.
Here's how the thinking goes. A boss like mine in 2002 could calibrate friction on purpose because the friction was visible. He watched me build the deck. He watched me walk into that room. If I'd flinched, he'd have known. Today, the exact same stretch assignment can go out to someone on your team, and they can quietly let AI walk them through the hardest part of it before it ever reaches you. The deliverable looks identical to what I turned in at Sony, and yet the person who produced it never had the experience I had. And you, watching from the leader's chair, have no way of knowing the difference, because the struggle that used to be visible now happens privately, in a chat window, before anyone else sees a draft.
The struggle that used to be visible now happens privately, in a chat window, before anyone else sees a draft.
This is where identity threat stops being a private cost and becomes an organizational one. A leader who avoids AI because of what it threatens in them isn't just falling behind on a skill. They're forfeiting the one vantage point that would let them recognize friction-avoidance when it shows up in someone else's work, because they've never felt the pull themselves. Shane Parrish, paraphrasing Richard Feynman, put this more plainly than I can: our biggest blind spot tends to be knowing our own weaknesses, because we are, all of us, the easiest people to fool. Part of why, he goes on to say, is that seeing our own flaws bruises the ego. That's Elul's whole argument in one phrase, from a book that has nothing to do with AI and everything to do with thinking clearly and making good decisions. A leader who has never caught themselves reaching for AI to skip a hard conversation has no internal alarm for the same move in someone they manage.
A leader who has never caught themselves reaching for AI to skip a hard conversation has no internal alarm for the same move in someone they manage.
The tell is subtle. A response that arrived too cleanly. An argument with no visible rough edges. You only know to look for it once you've noticed it in your own drafts.
I don't think the answer is refusing AI to preserve some romantic idea of struggle. Shirky is careful to say the same thing about students: opting out individually just creates a subculture, the AI-era version of bragging that you don't own a television. The answer has to be fluency, not abstinence, because fluency is what lets a leader tell the difference between AI used to build judgment and AI used to protect an identity that doesn't want to be tested. Somebody has to still be calibrating the difficulty on purpose. That was never going to be the technology's job. It was always going to be ours.
A year ago I asked what you needed to unbecome. This year, on the day set aside for reckoning with what your avoidance cost other people, I'd ask something closer to home. Who on your team got the stretch assignment this year, and do you actually know whether they sat in the hard part, or whether something else sat in it for them?
Break a Pencil,