The machine learning community (or AI more broadly) has collected truly extraordinary trophies in its very short existence of half a century a. The holy grail of superhuman performance in some games has arguably been achieved. Chess was conquered by DeepBlue in the '90s. More recently, AlphaZero by Deepmind showed us that machines can learn from scratch without explicit enumeration of rules. The success of AlphaGo is exemplar in this regard - we've beaten the world champion convincingly. More recent results with GPT-3 add to the growing body of evidence that modern approaches can be scalable although, precise understanding still eludes us. My examples are rather biased. Extraordinary achievements, however, are spread across all science. They've achieved things that no previous generations could even imagine.
Without discounting the extraordinary achievements, these results are a natural consequence of scientific deliberation. For the laymen, these results of momentary awe often become embedded in their daily lives to the extent that we take them for granted. Thankfully, extraordinary results usually live far longer within the scientific discourse to allow for deeper investigation before inevitably meeting the same fate. This fate, nonetheless, is a cause for celebration. It is precisely this stature that every idea (and the scientist) hopes for.
Contrary to the fate of scientific discourse, superstition has captured the imagination of people from time immemorial and still remains a vital component of many societies today. I conjecture that scientific discourse and superstition are more related to each other than we think and to an extent, might as well be the same thing. Both aspire to provide humans the ability to know and even influence the future. Both communities thrive on incomplete information, albeit in rather different ways. One uses it to fuel investigations of truth while the other only uses it as an engine of awe. Consequently, their fates are determined by this characteristic. While one seeks to amass experimental evidence to invent or discover natural laws, the other points to authority of yesteryears as irrefutable evidence. For one welcomes change, the other shuns deviation from the norm b. One believes in action and the other relies on the act of belief itself.
Given that the (meta-)objectives of both communities are the same, it begs the question of unifying the forces behind these two. Since unification is virtually impossible given the fundamental contradiction in their theses, science supplanting superstition is perhaps the more pragmatic goal. This thought is not remotely novel. Every moderately superstitious person, part of the overwhelming majority, would agree in a casual conversation that science should trump superstition. Why is it that despite convincing evidence in favor of science, superstition holds so many hostage? Understanding the answer to this question is precluded by our sense of the context in which ideas permeate societies.
A recent quote from a heartwarming post has stirred me,
With or without religion, good people can behave well and bad people can do evil; but for good people to do evil — that takes religion. - Steven Weinberg
In the absence of relevant context, no amount of proselytism can effect change. No facts, overwhelming evidence, and range of convincing arguments can deter the proponents of a potentially false theory. And the context lies herein - there hasn't been a convincing argument to the need of supplanting superstition. Barring small cults, everyday superstition remains innocuous to the level of being non-existent. I, however, believe that this is a boiling frog situation - while everyday acts of superstition remain harmless, they perpetuate a cognitive lethargy that hinders progress. It manipulates the layman to accept things for how they are. Awe trumps curiosity and contentment precludes action.
The key question then remains - how do you convince a frog that the water is really boiling?