The End of Theory - Is the Scientific Method Obsolete?

On the struggle of faith between dataism and humanity.

A New Holy Scripture

hris Anderson’s article The End of Theory as editor-in-chief of Wired magazine in 2008 is nothing short of a groundbreaking manifesto for the new religion of the 21st century: the data religion. In less than 1,500 words, he quite provocatively calls for replacing every scientific method from philosophy to physics with computational methods based on data.

While Anderson wrote this bold statement at a time when the first-generation iPhone lacked an App Store and ran on 2G connectivity, his argumentation is even more relevant today than it was in 2008. The data religion has attracted many followers since, while artificial intelligence and big data are about to transform almost every area of society, as well as science. Anderson accurately observed the technological trends and the enormous potential of rapidly improving computational power and ever-increasing datasets. But the intellectual appeal of his radical proposal is also the manifesto’s major weakness. As with any fundamental religion, the data religion needs to be challenged and scrutinised. Anderson’s proposal to replace science with statistics is at least highly questionable. But the underlying doctrines of the data religion appear to be quintessentially opposed to many ideals of humanism. The new holy scripture could therefore pose a danger to humanity itself.

True Prophecies

Anderson at times appears to be quite a visionary prophet. Many of his predictions, which must have appeared quite daring at the time, have retrospectively turned out to be highly accurate. While he noted a lack of organisational analogies for the petabyte age, Cisco estimates that we have since then not only surpassed the exabyte age, but have entered the zettabyte age, with an annual internet traffic of two zettabytes in 2020. The “most measured age in history” is being measured ever more accurately, “the biggest computing clusters the world has ever seen” are being clustered ever more tightly, and Google has applied mathematics and statistics in diverse fields from smart home technology and autonomous driving to energy efficiency and disease prediction.

More importantly, the data religion or dataism has since gained momentum and many devoted followers. Anderson describes a world “where massive amounts of data and applied mathematics replace every other tool that might be brought to bear”. A year later, American geopolitical strategist and forecaster George Friedman proclaimed the computer age and noted that “the computer represents both a radical departure from previous technology and a new way of looking at reason.” Some of the most firm dataism believers are the ones driving the digital revolution in Silicon Valley, such as Google’s Jared Cohen and Eric Schmidt praising that “technology is the continuation of evolution by other means”. Israeli history professor Yuval Noah Harari has stated that dataism might be the “single overarching theory that unifies all the scientific disciplines from musicology through economics to biology”.

Chris Anderson’s editorial has given readers an advanced understanding of the new possibilities of The Data Deluge in 2008, at a time when terms such as big data or machine learning were far from scientific mainstream. In this respect, he lived up to Wired magazine’s objective to be “the essential source of information and ideas that make sense of a world in constant transformation”. But his manifesto had a much wider impact. As a long-term author of the scientific journals Nature and Science, The Economist and several best-selling books, his theses attracted a larger audience of academics and practitioners, and correspondingly The End of Theory and the data religion provoked a variety of critical responses and counter-statements.

Questionable Preaching

The central postulation of Chris Anderson’s manifesto is that scientific models, being confronted with the unquestionable power of computer science and big data, will become obsolete as they are per definition fundamentally flawed and simplified, while statistical algorithms can find more accurate correlations based on almost infinite datasets. The ancient approach to science with hypothesising, modelling and testing will soon be regarded as outdated and obsolete.

The Harvard Business Review directly opposes this postulate in its counter-manifesto Why Data Will Never Replace Thinking. Data-driven methods have progressed science long before big data, and “the element of hypotheses/prediction remains important, not just to science but to the pursuit of knowledge in general”.

Big data might well be one of the most powerful tools science ever had at its disposal, but the scientific models that Anderson would readily dispose of will remain a crucial element of knowledge creation: “it doesn’t work without some thinking first”. It also remains an open question whether statistical analysis can come up with truly new scientific knowledge or is just relying on the scientific status quo. Harvard psychology professor Steven Pinker makes the case for a collaboration between dataism and scientific theories, in particular the humanities: “a consilience with [computer] science offers the humanities many possibilities for new insight”. The data religion might not be entirely wrong in its preaching, but the radical call for an abolition of all scientific methods remains highly questionable. It is in collaboration with the scientific method where dataism truly thrives.

The power of this collaboration can be seen in the way physicians will soon be able to fight cancer cells in the human body. Major advances in the sciences of molecular biology, genomics and immunology have decrypted the genetic signature of specific cancer forms. Harvard’s David Sinclair describes how, aided by computer algorithms and large datasets, physicians will soon be able to offer custom CAR T-cell therapy based on DNA-sequencing, even before the cancer cells develop in the first place. It might be interesting for dataism devotees to learn that the medical research teams who decrypted the cancer genomes worked with the ancient approach of hypothesis, model and test. Highly complex disciplines such cancer research profit immensely from new technologies, but the scientific method remains a crucial part of the problem-solving process.

Other scholars have further criticised the data paradigm. Mark Graham of the Oxford Internet Institute illustrates the biased and skewed nature of big data that is used to gain insight into social or political trends. He regards scientific models and human reasoning as crucial because “there will always be biases in how information and technology are used and produced”. Professor of digital humanities David Berry calls for the need to critically reflect on the algorithmic approach to reasoning: “it is problematic to erect an abstract and metaphysical standard by which human action and society can be judged “. Further criticism of radical dataism must surely include the ethical and practical limitations of data collection, the oligopolistic structure of the technology companies collecting the data, and the security and surveillance concerns about the governments and intelligence services using them. Being faced with such a multitude of concerns, science and the scientific method will probably not become obsolete any time soon.

Dangerous Doctrines

Even more disconcerting than the above-mentioned criticism and concerns about radical dataism are its underlying doctrines. Without delving too deep into details about the philosophy of singularity, it is certainly no overstatement that for the data religion not only science, but humans themselves might become obsolete as they are faced with an overwhelming capacity of computer algorithms. For radical dataists, the human organism is nothing more than a very advanced processor of information. Of course, as soon as the biochemical algorithms of the human brain get outperformed by electronic algorithms, there is no apparent reason why we should still rely on human knowledge and intelligence.

This can in many ways be regarded as a direct confrontation with the ideas and principles of humanism. German philosopher Richard David Precht regards this dangerous doctrine as a “cult of the inhuman” and “the triumph of the unreal over the real”. From a different perspective, Harari notes that we might very well “discover that organisms aren’t algorithms after all”, and it is “equally doubtful whether life boils down to mere decision making”. It remains to be hoped that in their quest for measuring and computing the world, dataism believers are not missing some essential parts of what life (and science) is all about.

A New Struggle of Faith

Like any other religion, the data religion faces criticism of faith from non- believers. While Chris Anderson’s technological predictions turned out more than accurate, his radical conclusion of abandoning all scientific methods except computational statistics remains highly challengeable. Instead, a collaboration of data science with other sciences from philosophy to physics might lead to previously unimaginable discoveries. On a more fundamental level, the struggle of faith between dataism and humanism comes down to a deeply philosophical matter: are humans truly unique from other (non-human) organisms, and if not, what is our raison d’être? It is to be expected that followers of dataism like Chris Anderson will become more vocal in their postulations as technology advances. Hopefully, articles like The End of Theory makes humanity contemplate more on one of the central philosophical questions of our time.

I write about the intersection between politics, business, and technology. Graduate student @Columbia @SciencesPo Affiliate @TheFutureSociety

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store