I was on a meeting with our team this week, and one of our people was visibly frustrated. Genuinely worked up. The problem: a complex component they were building was taking over seven minutes to deploy.
I asked the question I always ask when I see that face: "What's wrong — why are you mad?"
The answer: "It used to take five minutes."
I sat with that for a second, and it really made me think — because I could remember, very specifically, what that same process cost us before. A year ago, it took seven weeks. Three years ago, it took seven months. Months → weeks → minutes. And here we were, upset about two of the minutes.
Your sense of normal compounds as fast as your capability
Nobody warns you about this part of AI transformation. Everyone talks about the technology curve; nobody talks about the expectation curve that shadows it. Human beings normalize any baseline in a matter of weeks. Frustration doesn't calibrate against history — it calibrates against Tuesday. Give a team a five-minute deploy, and by Friday, seven minutes is an outrage. The seven months it used to take isn't even a memory; it's trivia.
Here's the thing though — and this is where most leadership takes would go wrong — the frustration is not the problem. The frustration is the asset. That team member's standards rose exactly as fast as the system we built. I don't want people who think seven minutes is fine because it used to be seven months. Rising expectations are how the next order of magnitude gets demanded into existence.
So what needs catching?
The emotion — before it makes the decision for you. That's the lesson I took out of that meeting: once you are fully AI enabled, you have to catch your emotions. Uncaught, the recalibration distorts everything:
- A two-minute regression gets treated like a production crisis, and someone burns a day chasing what a trendline would have told them to ignore.
- The team's morale runs on this week's friction instead of the three-year miracle they're standing on — you can be winning by orders of magnitude and feel like you're losing.
- Leadership overcorrects to the loudest frustration instead of the largest opportunity, because frustration is vivid and trendlines are quiet.
Catching the emotion doesn't mean suppressing it. It means doing what that question did in the meeting: naming the baseline out loud. "It used to take five minutes" — and before that, seven weeks, and before that, seven months. Once the whole line is visible, you can be two things at once: grateful for the curve and ruthless about the next two minutes. Both can be true — that's becoming the theme of this whole publication.
The rituals we run so the trendline stays visible
- Write the old baselines down. We keep the befores — the seven months, the seven weeks — on the record, the same way I keep predictions on the record. Memory is too kind to the past and too cruel to the present.
- Celebrate trendlines, not increments. Quarterly, we look at what a process cost a year ago versus today. The room goes quiet every time.
- Triage regressions by the line, not the feeling. Seven minutes against a five-minute baseline is a 40% regression and worth a look. Against the three-year line, it's a rounding error. Both facts get a vote; the second one usually wins.
- Ask the question. "What's wrong — why are you mad?" asked with real curiosity, is the cheapest diagnostic tool in leadership. It surfaced the entire recalibration in one sentence.
Why this is the best problem we've ever had
Think about what that frustration actually proves. The velocity that made it possible is an architecture decision compounding exactly as designed. The tiny team running it has standards so high that minutes offend them. And the future arrived so thoroughly that my team already takes it for granted — which is, if you think about it, the most complete possible proof of the rebuild.
Seven minutes is the new seven weeks. Catch the emotion, remember the months, and then — by all means — go get those two minutes back.
Clay, nine years on our team, put the same recalibration in his own words in his entry for the book: "the kind of work that didn't exist as a concept in our industry two years ago is now just what we do on a Tuesday." That sentence is what winning this transition feels like from the inside — the miracle becoming the baseline, over and over, until your only complaint is the last seven minutes.