Some days ago in the evening, after I finished speaking at a meetup in Dubai, two young guys waited until most people had left the room. They were not interested in debating Kafka internals or LLM benchmarks. They asked something much more personal, and much more relevant.
Where should we put our effort? What will be the next big thing?
Should we scale vertical or horizontal?
It is a question almost everyone asks at some point. It sounds strategic. It sounds ambitious. It sounds like you are planning ahead.
But the longer I work in technology, the more I believe that the question has the right intention, but it is maybe the wrong question to get the answer you really need.
The Seduction of the “Next Big Thing”
Five years ago, Machine Learning was widely seen as the ultimate career accelerator. If you were not doing ML, you were already behind. Then came LLMs, and suddenly everyone needed to “pivot into AI.” Today, answering emails with ChatGPT is enough to call yourself an AI enthusiast. Add a parallel subscription to Gemini or Claude and the label upgrades to “AI practitioner.”
More recently, the excitement has shifted again toward Agentic AI, autonomous systems, orchestration layers, and whatever acronym is currently circulating on conference slides. The pattern is familiar: each wave lowers the entry barrier for participation while raising the volume of self-identification.
At the same time, independent reports have shown that automated traffic on the internet already rivals or even exceeds human-generated traffic. In some recent bot traffic studies, more than half of total web traffic was attributed to automated agents rather than people. In other words, we are rapidly building a world in which software increasingly interacts with software.
The irony is hard to miss: while the terminology keeps escalating, the underlying mechanics remain variations of optimization, inference, control logic, and distributed coordination.
There will always be a next wave. As long as there is marketing and sales, there will be a narrative about the next revolution. Hype is not accidental. It is fuel. It attracts capital, headlines, attention, and talent.
But hype has one inconvenient property: nobody knows how long it lasts, how deep it actually goes, and how it behaves under real-world constraints like budgets, legacy systems, regulation, or plain human resistance.
And while everyone is rushing toward the next frontier, somewhere in the background there are still COBOL developers maintaining critical banking systems. You rarely hear about them. They do not post threads about “legacy-first architectures.” They do not have personal brands built around mainframes. Most of them are not on Instagram or LinkedIn explaining how they “pivoted into financial resilience engineering.”
And yet, if they stopped showing up tomorrow, a significant part of the global financial infrastructure would simply stop working.
The same is true for the engineers maintaining core networking stacks, the teams running payment clearing systems, or the people keeping airline reservation platforms alive that were written decades ago. They are not trendy. They are not keynote speakers at AI conferences. But their systems move money, route aircraft, and clear transactions every single day.
We do not notice them precisely because they work. Trend visibility and systemic importance are not the same thing. In fact, they are often inversely correlated.
Technology changes fast. Infrastructure changes slower. Fundamentals barely change at all.
If you build your career on waves, you are forced to paddle constantly. If you build it on bedrock (not meaning the AWS Service here), you can move between waves.
Vertical or Horizontal? That’s the Wrong Axis.
One of the two asked whether it makes more sense to grow vertically or horizontally. The question implies a trade-off: go deep in one topic, or stay broad and adaptable. On paper, my own career looks horizontal. I have worked in insurance, banking, retail, aviation. I moved from data engineering to streaming to AI. I speak about real-time systems and GenAI in the same week.
But if you take a close look at it, it has been very vertical. The constant has never been the industry, but the fundamentals. I love to understand the Algorithms and think about the mathematical reasoning. Understanding how distributed systems principles work out in real live and get a clue what actually happens under the hood.
If you understand those layers, moving from one industry to another is not a reinvention of your identity. It is more like learning a ne language.
The grammar of optimization, the logic of feedback loops, the mathematics of prediction and classification do not fundamentally change when you switch from retail to aviation. One plus one is still two. You can version-control it, containerize it, microservice it, and deploy it to three clouds in parallel. It will still be two.
The paradox is that the deeper you go, the more flexible you become.
Why I Told Them Not to Follow Trends
I also told them something that might make parents slightly uncomfortable: do not follow the trend.
Not because I believe trends are useless. Trends will always signal where investment and experimentation happen. But if you make them your compass, you will constantly feel behind. By the time you „arrive“ at one big thing, the narrative has already moved on.
Instead, I suggested something less glamorous: understand the mechanics. Learn the math and study the algorithms to understand why something works, not just learn how to call the API. Yes, this path is slower and it is more uncomfortable. It requires reading things, trying them out until you break them and that do not immediately translate into LinkedIn posts. It demands that you sit with confusion long enough for it clicks.
But in the end, if you understand how gradient descent works, you are not limited to one framework. If you understand distributed consistency models, you are not limited to one vendor. If you understand probabilistic reasoning, you can adapt when the tooling changes.
And tooling will always changes.
The Itchy Pullover Rule
There is another dimension that is often ignored in career strategy conversations: enjoyment.
You cannot excel at something you fundamentally dislike. You might be able to survive it. You might even perform decently. But you will never outgrow someone who is genuinely curious about the topic.
I once compared it to wearing an itchy but “nice looking” pullover to a party. From the outside, it looks great. Inside, you are uncomfortable the whole evening. You are distracted. You want to leave early.
Why would you build a career like that?
If you are naturally drawn to algorithms, go deep there. If you are excited by system design, pursue that. If you love product thinking, explore it seriously. Excellence is rarely forced. It is usually the byproduct of sustained curiosity.
Spot Problems, Not Buzzwords
Looking back, none of the major decisions in my career were driven by trends.
In the early 2000s, I saw growing amounts of data and organizations struggling with complexity. That led me to think about processes and modeling. Later, I observed that many business decisions were based on outdated data, which pushed me into stream processing long before it became fashionable. Over time, I became interested in what it means to be a truly data-driven company.
These were not reactions to headlines. They were reactions to observable friction. If you want orientation in a noisy world, start by asking better questions.
What is too slow?
What is inefficient?
Where are decisions made blindly?
Where does information decay before it creates value?
Trends often emerge because someone identified a structural problem and built something to address it. If you train yourself to see the problems first, you are less dependent on whatever label the market currently uses.
Titles and Salaries Are Weak Signals
Young professionals often overestimate the importance of titles. Over the years, the industry has become remarkably creative in inventing new ones. Senior, Staff, Principal, AI Engineer, GenAI Specialist, Architect. The vocabulary keeps expanding, sometimes faster than the underlying responsibility.
Titles can be useful for structure, but they are not substance. The real question is whether you are trusted with meaningful problems. Whether you have the autonomy to design solutions. Whether you are growing in your ability to reason under constraints.
I never chased titles, and I was rarely motivated by a marginal salary increase. My first filter was always the challenge. Is it interesting? Does it stretch me? Do I have enough freedom to solve it properly?
In the long run, that question has proven more durable than any label attached to a business card.
So Where Should You Put Your Effort?
If I try to distill what I told them into something practical, it is less about picking the right trend and more about choosing the right foundation. Invest in fundamentals that survive across industries and across hype cycles. Focus on problems that genuinely exist instead of buzzwords that momentarily trend. Choose areas that you enjoy deeply enough to push through the inevitable hard parts, because mastery requires more stamina than most people anticipate.
It also helps to separate visibility from depth. What is loud is not automatically important, and what is important is rarely loud.
And perhaps most importantly, accept that there will always be a new “big thing.” That is not a flaw of the industry. It is how innovation, marketing, and capital interact. The mistake is not that waves exist. The mistake is building your entire identity on predicting them correctly.
Career leverage does not come from catching the next wave early. It comes from building competence that remains valuable even when the wave has passed. Once you do that, the question of vertical versus horizontal growth becomes less dramatic. You develop depth where it truly matters and apply it broadly where it creates impact.
It may not be as thrilling as chasing every revolution.
But it is far more sustainable.