How Will Effort Be Measured in the AI Era?
Extensive use of AI and creation of numerous automations can be signs of many attempts. However, whether this actually reflects deep effort or sound judgment requires separate verification.
We can now have dozens of conversations with AI in a single day. We can use tokens in the thousands and tens of thousands, create multiple automation pipelines, and produce results quickly. Visible metrics accumulate much more easily and in much greater quantity than before.
This naturally raises a question.
Is using many tokens evidence of effort? Does creating many automations mean we're doing our work well? Does generating many outputs mean productivity has increased?
Metrics That Look Like Effort
In the AI era, metrics that look like effort have increased. And each of these metrics can have its own meaning.
Using many tokens can be a trace of conducting many experiments. It can mean not just asking a question once and stopping, but repeating the same question from multiple angles, attempting different approaches, or trying to find better expressions. Having created multiple automations can be the result of a reasonable attempt to reduce repetitive work. Generating many outputs can be evidence of not being afraid to execute.
These possibilities actually exist. It's difficult to deny them.
But the question doesn't end here.
Things Usage Alone Does Not Prove
The fact that many tokens were used is recorded. However, what that usage made clearer must be verified separately.
Repeating the same question with slight variations increases conversation volume. Yet whether that repetition made the question itself more precise, or whether it was merely confirming the same misconception in different sentences, cannot be determined by usage alone. AI returns answers regardless of how you ask. So asking more questions does not mean your direction was correct.
Automation is similar. Automation can reduce repetition. Well-designed automation allows people to focus on more important decisions. But if you chose the wrong things to automate, that automation merely produces incorrect repetition faster. An increase in the number of automations does not guarantee you are moving in a better direction.
The same applies to output. Producing a lot of output is a trace of execution. However, whether that output influenced subsequent decisions, or whether it made the next action clearer, must be verified independently of the quantity of output.
Using a lot is not the same as using well.
What Has Become Clearer
If we need a metric to measure effort in the AI era, we should add one alongside usage.
What has become clearer after that usage.
Have questions become more precise. Can you express the same problem better than before. Have standards become more explicit. Has judgment become faster or improved. Has the next action become clearer.
These things are not easily visible in numbers. They do not appear on AI dashboards. But if something remains after using AI, it is probably closer to these.
The direction in which usage accumulates and the direction in which understanding deepens are not necessarily the same. Sometimes the two move together, and sometimes only one moves.
Effort in the AI era is not about negating usage. It is closer to checking what usage has left behind.
Am I using AI a lot, or am I checking what has become clearer after using AI.