There have been two moments in my career when I genuinely felt I was standing at the edge of something infinite.
The first was in the late 1990s. The web was young, the tools were primitive, and almost none of the problems had been solved yet. You could build something over a weekend, put it online, and it would reach people across the world. The sense of possibility was electric. Every morning I woke up wondering what we'd figure out today.
The second moment is right now.
That Feeling, Again
After twenty-five years of building things — billing systems, satellite data platforms, mobile apps, maritime software — I have the same feeling I had in 1996: that the ground is shifting, that the rules are being rewritten, and that the people who pay close attention right now will look back at this period the way early web developers look back at the nineties.
The trigger? Working seriously with AI development tools. Not the first wave — the autocomplete-on-steroids generation. The current one.
The first time I had a real conversation with Claude about a complex architectural problem — not "write this code" but a genuine back-and-forth about design tradeoffs, edge cases, and long-term maintainability — something shifted. It felt like the pair programming sessions I'd dreamed about for decades: a collaborator who knew as much as I did, was never impatient, and kept finding things I'd missed. My buddy Hans (@hansv) was good for pair programming, and we exchanged ideas many times, but this AI... he was forgiving, patient, I dared ask questions that I wouldn't admit to Hans I didn't know the answer to. Explain it again. Simpler. Give me another example...
I thought: this is the late nineties, and I've been handed a browser for the first time.
Why Experience Is an Unfair Advantage
Here's something that took me a while to articulate, but which I now believe firmly: AI tools give experienced developers a disproportionate, compounding advantage.
This might seem counterintuitive. Shouldn't AI level the playing field? In some narrow sense, yes — a junior developer can now generate plausible code for problems they couldn't have touched before. But the analogy breaks down immediately, because the real work of software development was never just generating code.
The real work is:
- Knowing which problem is actually worth solving
- Recognising when an "elegant" solution will become a maintenance nightmare
- Understanding why the previous team made the choices they did
- Knowing which edge cases matter and which are theoretical
- Reading AI output critically enough to catch confident mistakes
These are judgment skills. They come from years of being wrong in interesting ways. No amount of AI-assisted code generation substitutes for them — and now, they make you dramatically more effective at leveraging AI.
For juniors, the situation is more nuanced. AI can accelerate their learning — but only if they treat it as a teacher, not a shortcut. The developers who use AI to skip the understanding step will find that their foundations are hollow, and that hollow foundations collapse badly when the system gets complex.
The experience multiplier: Every year of domain knowledge you have is now worth more than it was before. Not less. AI amplifies judgment — and judgment is what experience builds.
What This Means for the Next Few Years
The 1990s analogy holds in another important way: most businesses and most developers haven't yet figured out what this shift means for them. Just as in 1998, when the web was clearly transformative but most companies were still treating it as a brochure — we're at that early stage with AI.
The developers who understand this now — who are building intuitions, developing workflows, learning where AI is reliably excellent and where it confidently fails — will be in an extraordinary position in three years.
I've been building software since before most of today's junior developers were born. I've lived through the client-server transition, the web transition, the mobile transition, the cloud transition. None of them felt quite like this.
This one feels like all of them at once.
Want to talk about what this means for your business or your team? I'd genuinely enjoy that conversation.
Written by Kenneth Himschoot and published on January 15, 2026.