What is Software 2.0?

So far, the programmers have developed software by writing their own code. We call this software 1.0. A set of rules were given, and the computer followed them. Known conditions were outlined, and they were hardcoded. The key point here is that the software engineers used to code the software.

Since the technology has evolved, the methodologies have changed. AI has taken over all the sectors and simplified work for humans. Andrej Karpathy, director of artificial intelligence and Autopilot Vision at Tesla, has said in an article that in the future, the software will write itself. This is termed Software 2.0.

“Neural networks represent the beginning of a fundamental shift in how we write software. They are Software 2.0.”

Andrej Karpathy

Confused? Let me explain. As we know that in software 1.0, programmers developed software by coding in programming languages, while in software 2.0, the software will write its own code with the help of AI. It will be made possible by the use of deep learning.

So what is deep learning? In deep learning, the tasks are divided into layers and are connected through nodes. It is inspired by a human neural network and that’s why it is also known as the artificial neural network.

Meta (formerly Facebook), is also playing its part in the development of software 2.0 culture. They are working on Kotlin’s compiler to enable differentiable programming, and tensor typing in it.

In a nutshell, software 2.0 is going to shift the software development methods to a different level. It is going to be data feeding instead of explicit programming.