Wisdom Work and Weird Hours

Knowledge used to be the moat. You knew the tax code, the codebase, the vendor landscape, and people paid you because they didn’t.

That moat is draining fast. LLMs have the knowledge trained in. Anyone with a subscription can pull a competent answer on almost any topic in seconds. Chip Conley called this in Harvard Business Review two years ago: as AI commoditizes knowledge, wisdom work becomes the new knowledge work.

Wisdom is a different thing than knowledge. Knowledge is knowing the answer. Wisdom is discernment, taste, experience, and judgment. It’s knowing which question matters and which project to kill.

Here’s the practical version: the most valuable skill right now is figuring out what not to do.

AI will happily generate ten plans, all of them plausible. The person who earns their keep is the one who can say which nine to kill, which one to run, and what result means we picked wrong. A model can’t give you that. It hasn’t sat through your failures.

There’s a second shift happening that I think is connected.

Nathaniel Whittemore covered an essay by Bethany Crystal on The AI Daily Brief. Her thesis: weirdness will be the best human advantage. As a child of the early internet, being strange and peculiar was her standard. Then the web grew up, optimized itself, and asked everyone to do the same. Now AI has reintroduced a playful, experimental way of working. The shape of her workday stopped looking like anyone else’s.

That matches my life exactly.

I’ve built and iterated on real web applications from the driver’s seat of my car, talking back and forth with AI over iMessage on the commute. Not taking notes for later. Actually shipping. When a spark of an idea hits, I can act on it from pretty much anywhere.

Darren Hickey, who runs the digital agency Fellowship, goes on walks with Claude. He gets his exercise and works through his ideas out loud at the same time. Ten years ago that was a guy talking to himself. Now it’s a working session.

The standardized 9-to-5 knowledge workday was built around where the knowledge lived: the office, and the smart person in the meeting room. When the knowledge lives in your pocket, the workday deforms around your actual life. Weird hours, weird places, better flows.

Now the caution, because I’m as deep in this stuff as anyone and I still get burned out on it.

Humans were not meant to be always on. If AI lets you work from the car and the couch at 11pm, it also never gives you a natural stopping point. The same tech that makes work weird and flexible will happily eat every quiet moment you have. That’s not a tooling problem. That’s a judgment problem, and it’s ours to solve, not the model’s.

I heard something on a recent Data in the D episode that stuck with me. Mary Wilkerson made the point that we aren’t meant to consume this much, we’re meant to wrestle with what we read. Reading a hard book slowly is not an inefficiency to be optimized away. The struggle is where your own thinking forms. If everything becomes a summary, you never have the time and space to form an original thought. You just have opinions about other people’s compressions.

That lands hard for me, because summaries are the thing AI is best at. The temptation is to summarize everything and call it learning. But wisdom, the thing that’s actually scarce now, doesn’t come from summaries. It comes from reps. From sitting with a problem long enough that it changes you.

So the picture I keep coming back to.

Knowledge is cheap now. Wisdom is the job, and mostly it’s the judgment to say no.

AI un-standardized work. Take the gift. Build from the car, walk with Claude, work when the spark hits.

But protect the struggle. Read the whole thing sometimes. Sit with the hard problem without asking for a summary. Step back and go slower on purpose, because the discernment that makes you valuable is built in exactly those moments.

The teams that win the next decade won’t be the ones with the most AI. They’ll be the ones whose people kept doing the slow human work that AI made optional.