Big data will change the way people solve problems

In Critique of Pure Reason, the philosopher Kant mentions that truth is divided into the analysis of truth and the synthesis of truth. In short, the analysis of truth can be deduced from logical arguments, and the comprehensive truth needs empirical evidence and external data to prove it. In the past, we used to solve the problem by means of analytic method, first we established the model and law, then we put forward the new model and law by logic.

Innovation Endeavors's Zavain Dar suggests that with the development of computer systems and networks, the rise of large data and the spread of APIs will change the way we solve problems. In the past, we deduced new discoveries by establishing a novel analytical model, and now it has become the solution to the same problem through the building of infrastructure and capabilities and by means of integration.

He cites, for example, that for Web rankings, Google's scientists do not derive a theorem from the relationship between pages and links to better search, but to predict what future users want to see by collecting and synthesizing previous clicks and linked data. And Amazon has not summed up what the law of E-commerce to guide the understanding of who will buy what, but through historical events to optimize future activities.

Both Google and Amazon are early examples of the problem-solving approach from analysis to integration, since the two companies ' products are based on data from the digital medium. Data creation, storage, and interfaces used to interact with data are digitized and automated.

But now, with the popularity of large data and the popularity of APIs, the pioneers of data science and infrastructure have been able to gain incremental, predictable, comprehensive innovation and insight through the creation of high throughput and low latency facilities, abandoning the leap-forward, but time-consuming, analytical approach of the past.

Dar suggests that there must be two foundations to apply the integrated approach to new areas:

1 The underlying data must be in digital form

2 The process from data to scientists and back to data from scientists must be automated

The whole process of input and output must be automated.

The potential significance of the shift from analytic methodology to comprehensive methodology is enormous. In the past, these methods are limited to computer science domains such as artificial intelligence and information acquisition, and after all digitization, we will no longer need the knowledge and even the philosophical burden of the pre-set structure and axiom, and the popularization of the infrastructure can accelerate the popularization of the comprehensive problem-solving.

Marc Andreessen has raised the idea that software eats the world, and Dar believes that this view should be revised to digital datasets + input/output automation mechanisms + software will nibble the world with a comprehensive methodological approach to problem solving. For more information, please see this article in VB.

(Responsible editor: The good of the Legacy)