Norbert Wiener
Control is a loop, not an order
Seeing the result beats issuing instructions
Cybernetics, 1948
Why the thing you carefully explained still comes back wrong. How Wiener defined control instead.
What actually happened?
Cybernetics defines control not as issuing an instruction but as using the actual result to correct the next move — the feedback loop. Wiener's example is reaching for a cup: the eyes keep reporting the gap between hand and cup, and the hand corrects as it travels. Cut that report and the movement immediately goes off, however well it was aimed to begin with. Machines and living things, he said, obey the same thing here.
An instruction with no report back isn't control
However clearly you explained it, without a channel carrying the actual result back you hold an intention, not control. Things come back wrong not because people don't listen but because the loop is cut and nobody can tell how far off it has drifted.
Brief it and wait for delivery, never seeing the thing in between, and what arrives will not be what you pictured.
The report has to arrive while it can still be changed
Feedback later than the correctable window is worthless. The value of a loop isn't whether it exists but how often it reports — weekly and daily differ by an order of magnitude in how much drift can still be pulled back.
Finding out at month end that the month went wrong means the month is already spent.
Practising alone runs on the same rule
Stuck on a skill is usually not a shortage of practice. It's the absence of any mechanism reporting the result back: you finish and don't know whether it was good or where it went wrong. More hours won't fix that. A report channel will.
Months of practice and no progress: first check whether anything tells you which move was wrong.
How do I use it today?
Where you are: what you asked for keeps coming back different, or you've practised something for months with no progress.
Ask first: how often does the actual result come back to me? Inside that interval, how much is still changeable?
Where it goes wrong: reading it as watch more closely — he means shorten the reporting cycle, not raise the intervention rate. They aren't the same, and the second usually crushes the loop itself.
Lines to keep
Control isn't issuing an order. It's correcting from the result.
It came back wrong because the loop was cut, not because nobody listened.
Feedback later than the correctable window is no feedback.