In his landmark book “The Structure of Scientific Revolutions”, Thomas S. Kuhn introduced the concept of the paradigm-shift, which is a fundamental change in the way that we view and approach a particular problem or field.
For there to be a shift, the new paradigm:
- explains previously inexplicable phenomena
- resolves inconsistencies in the old paradigm
- fundamentally changes the way we approach the field
Today, we are witnessing another shift in the computing paradigm with AI making it easier and faster to create code, automate tasks, and more – and improving itself at an accelerating pace. Like other paradigm-shifts in computing, it is also enabled by the predictable increase in computing power (Moore’s Law). In addition, there is a ‘stacking’ or multiplier-effect of AI-oriented computing capabilities – that may prove to scale or multiply in a similar way.
Will AI-oriented computing be a true paradigm-shift, or simply a continuation of the current paradigm?
There is a striking fundamental difference from the current paradigm of humans writing code to explicitly define the steps a computer should take, to a potentially new paradigm of algorithms and models that learn and make decisions based on data and objectives.
The AI-oriented computing paradigm can explain things that were previously inexplicable and it can resolve many of the inconsistencies and inefficiencies of the current paradigm.
Perhaps most importantly, AI has the potential to enable new applications and solve problems that were previously impossible or impractical to address using traditional methods.
More broadly, a true Kuhnian paradigm-shift is NOT just ‘new technology or breakthrough’, but must include:
- a change in the way we think about and approach computing as a whole
- a shift in the underlying assumptions and concepts that guide our understanding
My Take
The current trajectory of AI-oriented computing is approaching a new paradigm.
AI-oriented computing could fundamentally change the way that we think-about and approach what we currently call the ‘software design / development / maintenance’ process – and towards new-thinking and approaches that are not centered on ‘the code’, and the multitude of tasks and human-efforts around it, that we see today.
I anticipate the new paradigm will be increasingly interdisciplinary – where the definition, purpose, and process of what we call ‘software’ today, will become more inclusive across disciplines and domains, with an increasingly broad span of capabilities; from the most ephemeral / esoteric personal use such as, ‘l need a one-time app composed for what I am doing today’ – to running the most sensitive and mission-critical business functions such as ‘execute according to required policies and business objectives’.
What do you think?
Will AI-oriented computing be a new paradigm, or is it simply a continuation of the current paradigm?