Artificial Intelligence as the Next Level of Knowledge Abstraction
There was a time when knowledge could be transmitted only through oral tradition. If you wanted to learn something important, you had to find a living carrier of that knowledge - sometimes the only one on an entire continent. Learning was slow, local, and fragile: when the storyteller disappeared, the knowledge disappeared with them.
The invention of writing became the first major leap. We created a copy of the storyteller - stable, portable, and scalable. Knowledge no longer depended on a single person.
Then came libraries - collective memory at scale. Later, computers emerged: machines capable of executing deterministic logic and taking over fragments of decision‑making. The internet connected these machines and made information retrieval effortless. Each new layer of abstraction accelerated human development exponentially.
Today, no one searches for a “wise elder” to hear a unique piece of knowledge. It’s simply inefficient. We live in a world where Google replaced the local library, and digital systems have become an extension of our memory.
Artificial intelligence is the next level of abstraction - perhaps the highest we’ve ever built. It doesn’t replace humans, but it removes yet another layer of limitations: speed, accessibility, scalability, the ability to work with massive volumes of information.
And that’s why it’s especially strange - and sometimes even surreal - to see its significance denied by software engineers, the very people who have always been the vanguard of technological change. Those who understand better than anyone that every new abstraction is first resisted, then normalized, and eventually becomes indispensable.
And if we’re honest, only a tiny fraction of today’s developers can read a compiled binary, interpret raw processor instructions, or understand the meaning of bytecode. It’s an absolutely minimal subset of the profession. Text‑based programming will follow the same path: it will become a niche craft practiced by a small group of specialists - not because it’s “obsolete,” but because the abstraction level has risen again. AI is doing to code what compilers once did to assembly, and what assembly once did to machine instructions.
Denying AI today is like denying writing in the age of printing or denying the internet in the age of global communication. History shows that each new abstraction doesn’t take power away from us - it amplifies it.
We’re standing at the threshold of another acceleration. The question is no longer whether we “need” AI, but how we will use this new layer to build more, learn faster, and expand the boundaries of what’s possible.