So by default, people are correct to mindlessly parrot this, even if they don't understand when they're wrong or misapplying the maxim.
also the irony is amazing
There is plenty of literature in the neuroscience and psychology fields that show we are biased towards finding patterns where there are none. Evolutionarily speaking, this is easy to understand. Better to run away one extra time, than getting eaten by a lion. The latter is less likely to pass on their genes.
> People mindlessly parroting anything is not a great success of anything, especially not of science communication.
I respectfully disagree. Appreciating the subtleties of causality is beyond what you can expect from most lay people. Even many of my machine learning colleagues can get tripped up by this stuff. So even if it is simplistic, I still think it is useful that people at least understand that causality is difficult.
This statement makes no sense at all. Pattern recognition applies to any data. That's not even contested terminology. (besides, just redefine your processed data to be your new raw data, and you're back to square one...and in many cases it wouldn't even be clear how to precisely define unprocessed data)
If you want to criticise the evidence in the literature, then fair enough I guess...that's beyond me to defend.
In fact, ALL clinical trials for all medicine are designed to verify causation. It's that straightforward. It just happens to cost a lot of money. Think about that. Why are clinical trials much more expensive than say the experiment correlating grip strength and life span? What is the differentiating factor?
This is the key. If you don't understand why/how clinical trials verify causation and what exactly is the differentiator between experiments that only correlate things then you don't truly understand the dichotomy between correlation and causation.
Most people don't understand it. Even the programming crowd. I even met data scientists who don't know about it.