What matters now is your vision

What matters now is your vision

---

When I say AI in this post, I mean the current wave: LLMs, agents, coding assistants.

The professional ground is shifting, and that opens a huge opportunity: learning something new or stepping into a different field has never been this accessible. That changes what one person can pull off.

Because the advantage is no longer in knowing a whole lot about one single thing. It’s in the vision of where to go and the judgment to recognize what of what you do works and what doesn’t. You still have to put in the work on the fundamentals, but they’re within anyone’s reach now. What’s scarce is that vision and that judgment.

The way we learn changed

It’s not a detail, it’s the core shift. Three things happened almost at once.

Learning speed shot up. Today you grasp the essentials of almost any topic in a fraction of the time it used to take, at whatever depth you want.

Building speed too. A first version that used to take months, today you stand up in days. But let’s not kid ourselves: getting from there to a final, working version still takes work, and a lot of it, because that’s where all the real software engineering lives.

And the methodology is far better, because it adapts to you. The key is that now you talk to it in natural language, the everyday kind: you don’t have to learn a syntax or a new tool, you ask the way you’d ask a person. You’re no longer the one bending to fit rigid material. Now the material molds to you: to your way of seeing the world, your experience, your context. You want it with diagrams, with flows, with tables, with analogies to something you already get. You ask for it and you get it, as many times as you need, without it ever getting tired.

It used to be “I tell you how to learn and you learn.” Now it’s “you learn however your head works.” That’s the difference.

I lived it crossing into other areas

I always wanted to write a Linux kernel module, even just a simple hello world, only to understand a bit how it all works under the hood. For years it was one of those “someday” things, because the barrier felt enormous: you have to understand how the system is put together, its methodology, its way of doing things.

With AI’s help, I did it. And it’s not that I became a kernel expert, not even close. But I understood the basics of how a module works: how it loads, how it compiles, how it talks to the system. Enough for something that felt out of reach to become reachable, and to keep pulling the thread on my own.

And here’s the question that matters: was I not allowed to get into the kernel just because I was never a kernel dev? Can’t I learn about sales just because I’m technical? That idea, that everyone has to stay in their box, is exactly the one falling apart.

And it’s not just the technical side. The same happens crossing the other way: sales, finance, design, whatever it takes to move a project forward. It’s not about becoming the best in the world at each thing, it’s about being able to step into places you never used to and create value there.

And there’s a bonus to crossing: when someone from another discipline enters a new field, they bring a perspective the insiders already lost. They see things the expert doesn’t, because the expert has spent years solving it the only way they know works. That mix, people coming in from different origins, is where a ton of new ideas come from. It doesn’t get in the way: often it’s exactly what the field was missing.

Now, it’s not black and white either. Sticking to what you know still makes sense: in your own turf you create more value and faster, and that’s not replaceable. But it’s no longer the limitation it used to be. There’s a middle ground, and the paths are crossing. I think we’re all looking for that balance: being able to step into other areas without becoming a bottleneck or ending up causing the opposite effect.

So why doesn’t everyone do it?

Here’s what keeps turning in my head. If learning and building have never been this easy, why isn’t there an explosion of people crossing into new areas and creating things? Why do so many still not know about everything, when they could?

I see two reasons.

One is fear. Some people would rather stay in their own lane and become the expert there, convinced that “AI is going to end up breaking everything and we’ll have to redo what we did from scratch.” It’s their way of waiting to see what happens, instead of using the tool already in front of them.

The other is less obvious: lack of vision. The limit is no longer access to knowledge. And the tricky part is that this gap doesn’t feel like ignorance, it feels like certainty. It’s not that you know you’re missing something. It’s that you don’t even suspect there’s more. It’s the “you don’t know what you don’t know,” but from the inside it doesn’t feel like doubt, it feels like being right.

It’s like holding a top-of-the-line phone and using it only for calls. You don’t feel limited. To you, a phone is for calling, period. You don’t look for more because, in your head, it’s already complete. You’re not missing information: you’re missing the suspicion that another map exists.

That’s why the bottleneck moved from access to imagination. And that’s why many don’t move: it’s not that they can’t, it’s that they don’t feel they’re missing anything.

And there’s another idea that doesn’t help either: thinking the interesting problems are already solved, or that they’re the same as always. They’re not. New problems showed up, and since the bar rose, now we have to solve more complex things. The good news is we also have better tools to be up to it.

“You don’t know about this”

When you create something outside your area, it’s normal for the line to show up: “that’s not your thing.” And it goes every direction: that if you’re frontend you don’t get infrastructure, that if you’re data you don’t know ML, that if you’re ML you don’t write good code. Everyone with their own fence. And it has its logic: what took you years to master stops being an exclusive advantage.

But the boundaries between areas are falling. Today work is done differently everywhere, in backend, in frontend, in data, in infrastructure. Almost nobody does everything by hand anymore. The line between “what’s mine” and “what isn’t” moved for everyone. It even happens with CEOs: today many act more like a product manager, deep in the product detail, and sometimes even like a junior dev opening simple PRs to improve things, leaning on AI. Clinging to that line doesn’t stop the change. Meanwhile, the one who crosses it is building.

AI doesn’t know, it gives you context

AI doesn’t know everything, and on its own it won’t make you an expert at anything.

What it does is give you good context on almost anything. But you’re the one who has to digest that context, connect it with what you already know, and turn it into something new. That’s where the value is: not in consuming the answer, but in generating knowledge that didn’t exist before, and sometimes in building it with your own hands.

And there’s something else: a model, by how it’s built, tends to give you the average answer, the most common, what already exists. That works as a base, but the good part shows up when you take that base and think outside the box: you connect things nobody had connected, you mix what you learned in one area with another, and you come out with something that wasn’t there. AI gives you the starting point. The original part you put in.

What I do with this

I’d rather expand, learn whatever it takes, and build. It’s not the only way to do things, it’s the one that makes sense to me today.

The way I see it: knowledge is now within anyone’s reach. What matters now is the vision of what to build and the judgment to recognize what generates value and what doesn’t. The how, which used to be the barrier to entry, no longer is.