As a developer with an Industrial Engineer background, I have been developing a way to measure real productivity when using a user interface that is independent of the user’s skill or computer delays.
Traditionally, we could do a time-and-motion study and give you the savings in seconds, or we could do a predetermined motion time study (which means using tables to calculate the movements of the person to come up with a time). In the past, my go-to tool was MODAPTS, which was developed by Chris Heyde in 1966.
Today, I’ve decided there are easier ways to explain productivity gains by creating my own standard units called ‘moments’.
So moments are grouped into Knowledge moments, Procedure moments, Human moments, and AI moments.
A decision moment is when the human brain or an LLM must interpret a situation and trigger a procedure.
For example, shutting down a Windows 10 computer (not putting it to sleep) can take 3 knowledge moments and 3 procedure moments, totaling 6ʍ.
We can call the unit of a moment the inverted w symbol ʍ.
Whereas Ammenster’s total moments for a human to shut down the PC are 3ʍ, but the knowledge moments are only 1ʍ.
For an AI, it takes as little as 2ʍ, but realistically, much more, because in real-life situations, the AI may have a checklist to go through before shutting down.
In any case you can probably see that this simple analysis using moments ʍ that it is possible to objectively measure how many descisions a human or AI are making, thereby how much decsion load is on the process.
.
Leave a Reply