Pathology of Data science

“If Jupiter stands inside the moon, in this year the king will die; variant: there will be an eclipse of the moon and Sun; a great king will die. [—] If Jupiter enters the moon: there will be famine in the Westland; the King of Elam will fall in battle; in Subartu a noble will revolt against his lord. (Hunger 1992, 100: 60) - “ Assyrian Astrologer from second Millenia BC "

To know the future, is an urge that humans are innately born with . It is our apriori(Kant) that we can’t think outside the cause and effect framework that forces us to believe that everything must have a rational cause.We refuse to believe that our misfortunes are not in our control, we refuse to believe that unrelated large movements in our surroundings are going to impact us in ways that we cannot comprehend . Our inability to not reason about things that are out of our control cause insecurity to us ,which are further exploited by people who exactly know how to exploit this apriori urge. There is no much of a difference between the claims of Assyrian astrologers and data scientists of the modern day.

There is an entire brewing multi Billion dollar industry, running around this new found technological miracle that no-one seems to understand or at least bothers to , there is a Utopia, that it will predict future, it will predict what could go wrong and it will even predict what it cannot predict with specific determinism and accuracy of probability . The problem is, businesses were always out of control , business men were always vary of the undeserving monopoly they have and how they might loose it in a movement of ignorance. One is always insecure of what they don’t deserve. Astrologers, Economists (especially the macro kind), MBAs and Political Scientists( I don't know when that happened , but they rebranded themselves as scientists) and now data scientists , have always found creative ways to bait insecurities in people for their survival. We have always had this astrology driven economy , astrology never really died.

It’s easy to blame AI or ML , but it’s in the pathology of Capitalism to Automate things and predict future and stay in control of it. In short we are not automating things because AI happened, on the contrary AI happened because it’s in our capitalist pathological DNA to automate. To automate people, practices, governance, rules, systems and every thing that follows it . Capitalism is using Democratisation of AI as an excuse for its pathologies. If AI were to go down another winter , so be it , but businesses and monopolies will always find newer ways to commit suicide. Most organisations which have never had a strong tech team and are over 30 years old most likely will not see the face of the next decade(as in 2030) and that should be the best case scenario. The worst case scenario would be , the same firms hoarding up monopolies to destroy economies.

One might think I am being a bit pessimistic , how can a technologically advanced world fall victim to a bunch of data scientists? Well I would say astrology has survived for at least 4000 years now and the nexus of astrology + religion has devastated entire Civilisations and set back countries into eternal dark ages.We might think we are a technologically advanced civilisation , but that's exactly what the assyrians thought about themselves. As Bertand Russel would say “A process which led from the amoeba to man appeared to the philosophers to be obviously a progress though whether the amoeba would agree with this opinion is not known.” ~ Bertrand Russell”. Furthermore, Even well run academic disciplines are susceptible to corruption and entire generations loosing out on efforts , String theory in physics is a great example where things turned into a ponzy scheme.

At the start of every new era of technology , be it Stone Age ,Bronze Age humans have flourished until a balance was established and once all progress that could be made using the new tool has been made, immediately pseudo sciences and religions take over societies. Arresting all progress and pushing civilisations into dystopia. Now it’s hard to say if dark ages cause pseudoscience-friendly societies or its the other way round, but Popper in his Open society and its enemies strongly points out that it is always the pseudosciences first and then the down fall of the society.

For the first time again, in may be a 1000 years, we are democratising astrology (from this point I will use the words data-science and astrology interchangeably). I am not necessarily against the idea of democratisation of AI , not in the spirit atleast . Modern Sciences , since Galileo , to a large extent have remained under the control of academia and research organisations ,which strongly adhered to peer-review mechanisms and stuck to the first principles of Philosophy of science. Despite the best efforts even academia had its own dark spots . I am afraid , in the hands of wrong people (form this point on I will use the phrase “wrong people” and MBA’s interchangeably) entire economies can collapse, which may be a good thing in a well ran competitive environment. But the problem is , these wrong people who go to wrong places(You Guessed it right , Business Schools) to get their degrees and build strong monopolistic networks and turn Business schools into glorified contact books . This is invariably the root cause of the pathology of capitalism .These poor organisational principles combined with a dose of astrology , they will take entire organisations down with them.

Most organisations are not built to handle science , Popper’s philosophy of science is an essential tool that keeps academia from falling off the cliff. Most leaders don't bother about concepts like peer-review and falsifiability, these guys would value political correctness and mutual respect over peer-review and falsifiability anyday, especially when it involves them. I have certainly met a data science manager, who honestly and sincerely thinks LSTMS can be applied to predict which next job one will get hired, looking at the names of companies applied in the past (its just sequence modelling). No , that isn’t how it works , Language Machine learning models are designed to work on rules of grammar with a finite amount of context. There is no Turing machine that can learn a language of job-hiring. Because there is no such language that defines the grammar rules in this domain, Unless one doesn’t understand the Nuances of Theoretical Computer sciences, VC theory and other foundational philosophical intricacies of AI its very hard to not make a complete fool out of yourself. Had Popper been alive , he would be rolling on the floor laughing at the current state of affairs.

Now, AI as a field has always been dancing with pseudosciences . From Dreyfus to John Searle, there are enough critics of AI (Enough controversy for today , will not get into this . We will assume AI works, in the hands of right people). This is exactly the reason why Turing had always stayed away from making any unrealistic claims. His 1950 paper is a great introduction and a very well balanced starter for most .

Here is my advice for organisations:

• If you are a big old organisation , never been in tech and believe that there is a chance AI can take over your business : Sell it right away while you can and save some grace.
• If you are a startup with no strong Foundational CTO or Tech head who can drive, and want to use Astrology as a business model : Never been a better time , welcome to party . wait for 10 more years and then go to step 1.
• If you fall into neither of the above categories, hire your managers who have a strong understanding of Philosophy of science and has probably read Popper and Kuhn. It’s just not sufficient to be a good engineer or manager anymore in the era of democratisation of science.