First of all, I need to clarify something. Like many people, I use the term AI all the time. But what I’m really referring to is machine learning. Machines that learn.
And the more I think about it, the more I believe machine learning belongs in the same category as operating systems, web browsers, and sandboxes…not because it replaces them because it is becoming a new environment.
A New Layer
Think about what an operating system really does: It sits between the user and complexity. The user clicks a button. The operating system takes care of everything else. The user doesn’t care about memory allocation, interrupt handling, drivers, or file systems. They care about getting the job done.
Machine learning is starting to play the same role. It is becoming another layer between humans and technical complexity, a new API, a new environment, a new operating system of sorts.
Always On
What makes this operating system different is that you don’t even need to boot it 🙂 It’s already there, You don’t inytall it, you don’t patch it, you don’t spend hours configuring it…you simply interact with it. The interface is conversation and that is a massive shift.
For decades we learned how to communicate with computers. Now computers are learning how to communicate with us.
The End User Doesn’t Need To Know
Don’t get me wrong. This is not another one of those articles claiming that nobody needs technical skills anymore because AI can do everything. Far from it.
What I’m saying is that the average end user of a machine learning system does not necessarily need to understand the underlying mechanisms. They don’t need to know how models are trained. They don’t need to understand optimization algorithms. They don’t need to know what happens behind the curtain.
They simply need results. Just like most drivers have no idea how an automatic transmission works. Yet millions drive every day.
“The driver doesn’t need to build the engine to get to work.”
But Somebody Still Has To Know
This is where many people get confused. Higher abstraction does not eliminate expertise. It changes where expertise lives.
I remember the days when we had to get into CONFIG.SYS and AUTOEXEC.BAT. We had to understand memory management because the operating system expected us to.
Today, most users never think about memory allocation. Tomorrow, most users may never think about many technical processes that currently consume entire professions. That does not mean the mechanisms disappear. It means fewer people need to understand them. But the people who support, design, and build these systems must understand them better than ever. The operating system still needs engineers.
The New Support Desk
Every major operating system created a new generation of experts…DOS experts, Windows experts, network experts, system administrators…
The machine learning era is no different. A new operating system creates new support desks: New architects, new engineers, new specialists…
The average user becomes more productive. The expert becomes more important.
Welcome To Another Operating System
Maybe the biggest mistake we can make is treating machine learning as just another application. I don’t think it is. I think it is becoming a new computing environment. A new abstraction layer. A new operating system that sits between humans and complexity. Most people will simply use it. A smaller group will understand how it works. Just like operating systems before it.
And just like before, the people who understand the internals will be the ones building the future.
“The better the operating system, the less the user notices it exists.”