Over 3 years ago I started voice dumping a half-baked and formed idea into Google Drive and then asking GPT 3/4/5 to turn it into a blog post. Even though a lot of them turned out to be AI slop writing, I still find value in revisiting them from time to time. So with the new 5.6 Sol model, I decided to revise an earlier version of an idea that Goodhart's Law of "when a measure becomes a target, it ceases to be a good measure," is often applied to success with a correlation does not equal causation mistake.
[The following is a GPT 5.6 Sol revision of a GPT 4 post from August 2023]
A successful CEO wakes up at 5:00 every morning.
An aspiring entrepreneur hears this and sets the same alarm. The logic seems simple: successful people wake up early, so waking up early must help create success.
Maybe it does. The extra time could be used to exercise, plan, read, or work without interruption. But the CEO might also wake early because of young children, anxiety, a demanding schedule, or a natural preference for mornings.
Waking up early may be correlated with success without causing it.
That is the familiar warning that correlation does not equal causation. But something more interesting happens when we turn the correlation into a goal.
Once waking early becomes a visible marker of ambition, people begin optimizing for the marker. They post screenshots of their alarms, promote elaborate morning routines, and treat exhaustion as evidence of discipline.
The signal of success becomes the strategy for achieving it.
This is where correlation meets Goodhart’s Law:
When a measure becomes a target, it ceases to be a good measure.
Consider a school judged almost entirely by test scores. The scores may begin as useful measures of learning. But once they become the main target, the system adapts.
Teachers narrow the curriculum. Schools devote more time to test preparation. Resources shift toward students closest to passing. In extreme cases, results are manipulated.
The score still exists, but it no longer measures learning in the same way. Targeting the measure changes what the measure represents.
The same thing happens when we imitate successful people.
We notice behaviors associated with desirable outcomes:
Executives wake up early.
Writers read constantly.
Entrepreneurs take risks.
Productive people follow strict routines.
Happy people exercise or meditate.
We then assume that copying the visible behavior will reproduce the result.
But we may be copying a signal rather than a cause.
Take the advice to “take more risks.” Successful founders often did. But we mostly hear from the founders whose risks worked. Those who took similar risks and failed are less likely to write books or appear on podcasts.
The outcome may also depend on timing, skill, relationships, financial security, market conditions, experience, and luck.
Copying the risk without recreating those conditions is not copying the strategy. It is copying one visible part of a much larger system.
Celebrity diets work the same way. The public sees the meal plan. It may not see the trainer, chef, flexible schedule, medical supervision, genetics, or hours of exercise.
The easiest variable to observe becomes the easiest one to target.
This suggests a broader version of Goodhart’s Law:
When a correlate of success becomes a prescription for success, it becomes less likely to produce the success it once predicted.
As more people optimize for the signal, the signal becomes less meaningful.
Waking up at 5:00 may once have suggested discipline and purpose. Once everyone adopts the habit because successful people do it, waking early tells us much less. It may reflect discipline, or it may reflect imitation and sleep deprivation.
When learning from successful people, we often copy their practices when we should search for their principles.
The practice might be waking early. The principle may be protecting uninterrupted time.
The practice might be reading fifty books a year. The principle may be continuously improving your understanding.
The practice might be taking large risks. The principle may be accepting uncertainty when the potential upside justifies it.
Practices are visible and easy to imitate. Principles require interpretation. They force us to ask why something worked and whether the same conditions apply to us.
Much of modern self-improvement consists of dressing ourselves in the habits of successful people. We copy their schedules, diets, reading lists, software, clothes, and workspaces because doing so makes us feel closer to their results.
But success is not a costume.
Visible habits are often the surface expression of deeper qualities: judgment, persistence, skill, relationships, opportunity, and accumulated knowledge. Habits may support those qualities, but they cannot automatically create them.
Do not ask only:
What do successful people do?
Ask:
Why did it work for them, and what would have to be true for it to work for me?
Otherwise, we may spend our lives optimizing the signals of success while moving no closer to its causes.