Efficient vs Effective

Are you worried about how you can be more efficient with your work, especially with AI? That's a trap.

Cal Newport's podcast episode today (November 16, 2026) was pure candy for a nerd like me. He talked about the roots of our obsession with efficiency. It all goes back to measuring how well a machine converts one type of energy (potential energy for a water wheel, heat for a steam engine) into useful energy.

In that setting, you can think of efficiency as a fraction of how close your machine gets to the theoretically perfectideal. For modern audiences, the best example is a solar panel. A hypothetical panel that converts all the incoming solar energy into electricity would get an efficiency of 100%. Anything below that gets something less. Then we can compare different material designs, coatings, and arrangements and measure how it affects that ratio of produced energy to maximum possible energy.

Managers took that idea and ran with it, but in doing so made a category error. The ratio of things to compare is now neither dictated nor bounded by physics. A manager might choose to measure "forms filled per clerk per hour," but as that clunky phrasing suggests, it's not a simple ratio that tops out at "100%." There is no clear upper limit. There is also no clear indication that this is a number your business should measure or try to improve.

And that brings us to AI. It is clearly very efficient if we measure it with these simplistic ratios.

  • Number of outbound marketing messages sent per day? Through the roof!

  • Lines of code produced per day? In the thousands.

  • Meetings transcribed and summarized? All of them.

Great. But...

  • How many paying clients did all that outbound net you?

  • How many paying users did all that code bring in?

  • How many important decisions has your organization moved forward on?

Zoomed in all the way to the minute-by-minute execution, AI appears incredibly efficient. Whether, zoomed out to the strategic length of a few months, it has also been effective is an entirely different question, but it's the only question that matters.

To tune your business so it becomes an effective machine rather than a purely efficient one, be careful about where and how you deploy AI. Don't look for volume. Look for quality.

Next
Next

Make More Disposable Mini-Tools