One of my favorite reads this year has been Aruna Ranganathan and her student Xingqi “Maggie” Ye’s article in Harvard Business Review, “AI Doesn’t Reduce Work — It Intensifies It.” In case you’re paywalled, Berkeley Haas has a great write-up.
I seem to be experiencing this same phenomenon this weekend.
As I finished a draft of a paper I have long owed my coauthor, I started working rigorously on a project I haven’t touched in a very long time. I told Claude Code to do things, and oversaw its work. There was a flywheel — as Claude surfaced things, I shared thoughts, we moved forward, it worked, I had thoughts, those thoughts led to checks, we kept going. Claude and I have probably completed two weeks’ worth of work in two days, almost none of it writing.
I am exhausted right now. Thank God I have dinner plans, because I probably wouldn’t be able to stop otherwise. And I had so much fun. I got to direct — because I know, and have known, all this data work needed to be done — and sit back and look at the results. Enthusiastically — sure the effect I was finding on one variable is a wash, but the main story is turning out in spades. Passing every robustness check I throw at it. I feel like Mario racing past obstacles.
It’s fascinating — and I think it gets worse the more agentic we turn. New meanings to “work hard, play hard.”