The Day Productivity Stopped Meaning Anything

A few days ago, I was sitting in a booth at a cluttered bar with a friend (Himanshu), trying to talk shop over the music. We were complaining about the usual, performance reviews, the quarterly cycle, the pressure to "ship more."

Then he leaned in and said something that stopped the conversation dead.

"Maybe we shouldn't be reviewing performance anymore. Maybe we should be reviewing how people think. There ability to think critically"

I didn't agree immediately. I actually pushed back. I told him that thinking doesn't pay the bills; shipping does. We sat in silence for a second, watching the ice melt in our drinks. But the more I tried to dismiss it, the more it felt like he had just pointed out a crack in the foundation of everything we've built.

If you really sit with that thought, it quietly challenges the entire structure of how we have valued human effort for decades.

The Collapse of Scarcity

For the longest time, work was about cognitive muscle. Writing boilerplate, memorizing syntax, and implementing patterns from memory required time, skill, and energy. Because these tasks were difficult and slow, high output was a reliable proxy for high value.

But we just hit a wall. AI has removed the constraint of effort.

Last week, I watched a junior developer generate a complex authentication system in twenty minutes. It used to take me three days. When the cost of generation drops to near zero, the traditional metric of "how much did you ship?" becomes an empty data point. If a machine can do the "work" part of work, then "performance" is no longer a measure of the person. It is just a measure of how long they left the tab open.

The High-Speed Pipeline

The danger isn't that AI is bad. The danger is that it is too good at being average.

When we prioritise volume, we inadvertently turn engineers into pipelines rather than architects. You see it in the code, everything starts to look like a "beautiful Frankenstein." It's a system where every individual function is perfect, yet the whole thing feels wrong.

This happens because AI has no "why." It operates with a stateless intelligence. It doesn't remember that you chose a specific database last year because of a unique security requirement. It just gives you the global average. If a human doesn't intervene, you aren't building a product; you're just hosting a collection of statistically likely patterns.

From Execution to Supervised Direction

This is where the shift gets messy. We like to think we're moving from "doing" to "thinking," but that's too clean. The reality is more uncomfortable: we are moving from manual execution to supervised execution.

It's the difference between a chef and a factory supervisor.

I used to spend my days solving puzzles. Now, my value lies in my ability to be a "Director of Intent." This requires a specific kind of friction that we usually try to avoid in corporate settings:

• The Intentional Reject: The real work is now found in the moments where you say "no" to a perfectly functional suggestion because it doesn't fit the soul of the project.
• The Weight of Trade-offs: AI is terrible at trade-offs. It wants to give you everything. A human has to decide what to sacrifice. Experience is just compressed failure; it's the memory of the last time a "perfectly functional" solution crashed the legacy system at 3 AM.

Experience is the New Scarcity

This is the part that keeps me up. Learning has become nearly free. Anyone can ask an AI for a system design and get a reasonable-looking answer. But experience? Experience is still expensive.

Experience isn't just knowing the answer. It's the confidence to reject the popular answer.

A "learned" person sees the AI's output and thinks, This works. An "experienced" person sees it and asks, Why this approach? What happens when this hits ten thousand concurrent users? How does this align with who we want to be?

They don't treat the machine as an answer engine. They treat it as a collaborator that is constantly trying to take the easy way out.

The Uncomfortable Truth

We used to value people for how much work they could do. Now, we have to value them for how much unnecessary work they can prevent.

That is a much harder thing to put on a spreadsheet. It's a quieter, more invisible skill. It's the art of discernment, the ability to look at a sea of infinite, free output and know exactly which 1% actually matters.

If we keep measuring people by their volume, we will end up with mountains of perfectly executed noise. The real risk of the AI era isn't that we'll build things that are wrong. It's that we'll build things that are forgettable.

The shift is here: from performance to discernment. I'm still not entirely comfortable with what that means for our titles or our salaries, but I know one thing for certain. If you can't tell the difference between the two, the machine has already replaced you.