Perhaps you're just being incomplete in your final thought there, but I think what you've said as is really exemplifies the problem. Science is not just fitting something to data. Science is about predictions, testability, falsifiability, and more. Give any set of input/output data and you can create some sort of model or regression that fits to that data. But that model does not inherently mean so much, especially as the inputs start to become more complex -- meaning the output might not even be mapping to anything fundamental. I think the article hit most strongly on that with this comment:

"Hossenfelder’s real target, when you strip away some unfortunate terminology, is not beauty but self-satisfaction, which encourages disengagement from reality. That attitude reaches its theoretical apex in the doctrine of “postempirical science,” which argues that social consensus, not experimental evidence, determines scientific validity. Here she quotes physicist George Ellis, rebuking physicists and philosophers who adopt that attitude: “There are physicists now saying we don’t have to test their ideas because they are such good ideas. They’re saying—implicitly or explicitly—that they want to weaken the requirement that theories have to be tested. To my mind that’s a step backward by a thousand years.”"

Much of 'science' now a days has started to veer down the path of pseudoscience, and the path there is paved with regression mapping and 'science' by consensus.