Stay Connected
Sign up to receive our newsletter straight to your mailbox.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Something has shifted lately in tech, and it’s not just an uptick in innovation. Engineers are used to churn. Tools, frameworks, and languages all come and go, and staying relevant has always meant continuous learning. So obsolescence isn't shocking; it's just part of the profession.
What's different now is the rate of acceleration.
Engineers are being asked to do more with less — whether due to the end of the ZIRP era or AI already writing most of the code for many engineers. The half-life of purely technical advantage is shrinking. Technical expertise brought the paychecks and the prestige. But what if it's becoming less relevant?
As AI agents take over the drudgery, the people stuff is going to matter more, not less. As execution velocity increases, misalignment scales with it. Coordination becomes increasingly important.
As AI agents take over the drudgery, the people stuff matters more.
But the people stuff can be a real pain. Poorly run meetings have you itching to get back to “real” work. It's easier to keep your head down and work around the difficult people. Avoiding conflict while being consistently competent should be enough to earn respect. And when conflict becomes unavoidable, the best argument should win.
An engineering education — and most engineering careers — don't cultivate people skills. When technical expertise was the dominant variable, it was easy to treat all that as secondary.
For most of my career as an engineer, I operated this way. What eventually became clear — slowly, and only after failing in ways that were difficult to blame on anyone else — was that my limitation wasn't mainly technical. My problem was somewhere I wasn't looking.
Growing up, math and science were my favorite subjects in school, in part, because there were correct answers. If I got the answer right, I got the grade. The system made sense. Later, that same mindset translated cleanly into engineering: isolate the problem, fix the bug, improve the system.
For a long time, I didn't recognize how this logic shaped the way I approached relationships.
If someone disagreed, I tried to explain again, just better. If someone seemed upset, I tried to be more reasonable. I eventually noticed a pattern — I was subconsciously keeping score. If I was sufficiently kind, competent, or accommodating, then other people should figure out what I wanted and give it to me — without me having to ask. If I did the "right" thing, the "reward" should result.
It didn't.
Perhaps relationships just didn't make sense. They certainly weren't predictable — the same action produced different responses with different people, sometimes with the same person on different days. But they weren't random either. Patterns repeated. Misunderstandings resurfaced. The same dynamics showed up across different relationships, different contexts, different people.
What if relationships operated by a different logic entirely?
Your 1:1s are fine, but rote. The feedback you give isn't followed, and the feedback you get isn't fully accounting for your circumstances. A performance review catches you off guard — and you're left wondering if you missed something or if it was being withheld. A disagreement you thought resolved keeps resurfacing in slightly different forms. The latest culture initiative seems like just another set of slogans to espouse.
These problems should yield to obvious fixes. Someone wasn't clear enough. Someone didn't listen. Someone needs to adjust their behavior. Less talking. More coding.
That diagnosis feels right. And it keeps not working.
Perhaps you've noticed a particular dynamic resurfacing across your career — with different people, different projects, different teams. What looks like a personality issue repeats across personalities. What looks like a one-time breakdown keeps re-forming. You address what seems like the actual problem, and something structurally identical shows up later wearing a different face.
This might bring you to an unsatisfying conclusion. Either people are not skilled enough at the communication part, they're out to get you, or they're just driven by a random number generator. Whatever the case, it leads to the same place — try harder, or just give up.
There's a third option. The diagnosis is wrong.
You've got a successful strategy for technical issues — isolate the problem, apply the fix, test the result, iterate if needed. When a program misbehaves, you change the code, you run it again. The system can be paused, inspected, modified, and relaunched. You are operating the machine.
Relationships don't work like that.
You can't roll a relationship back to a clean state and relaunch it without history. What's done is done. There's no reboot, no root access to other people’s minds — and crucially, there is no separation between you and the system you're trying to fix. Much like the observer effect in physics, your diagnosis of the situation is itself a variable inside it. The tone you bring to the conversation, the certainty you project, the patience or frustration you carry — these don't just transmit information. They shift conditions.
Seeing this is like finally getting a punchline. In the Seinfeld episode, George Costanza — having spent a lifetime following his instincts straight into failure — arrives at a revelation: "If every instinct you have is wrong, then the opposite must be right." He tries it. It works. The episode plays it for absurdity, but the underlying insight is accurate. When a relationship settles into a pattern that isn’t working, doing what comes naturally just reinforces it. Doing the opposite interrupts it.
Some people figure this out strategically. "Act like you don't want it." "Play hard to get." "Withdraw and they'll lean in." These techniques work for the same reason George's experiment worked: tactical yielding reduces resistance, withdrawal triggers pursuit, and faux scarcity shifts perceived value. They're not “games” by accident — they're systems awareness deployed as leverage — unilateral moves with no concern for making the pattern visible or the dynamic mutual. Dark-side version of the same logic.
The light-side isn't just a game. It's fluency in working with interaction patterns. Relational skill can feel counterintuitive because it requires a shift from operating on a system to participating intelligently from within it. The leverage points don't follow simple cause-and-effect. It can mean pausing where you'd normally jump in. Making space for others' initiative even when you're confident in your own. Or taking up space where you'd normally wait-and-see. Instead of trying to win, you keep the exchange alive. Instead of figuring out who to blame, you own your part.
It's not manipulative or magical. It's just a different ruleset when the system is not separate from you.
The instinct to diagnose and fix from outside is the move that keeps you inside the pattern.
The workplace patterns described earlier aren't just communication failures. They're stable equilibria — configurations that persist because every move reinforces the next.
The person who smooths over tension keeps everyone comfortable, yet nothing resolves. Initiative-takers and wait-and-seers fall into the same two-step every time. Looking for someone to blame makes people act like they've got something to hide, even if they don't. And keeping a low profile can make it look like there is something to hide. The advocates and devil's advocates take their positions. Tactfully avoiding a conflict just ensures it stays unresolved until the next time it can be avoided.
Nobody agreed to any of this. It just became the way we do things.
These patterns hold not because of bad intentions, but because everyone is participating in a way that feels familiar — and even functional, up to a point.
Picture a moment in a room you know well, with people you know well. The conversation has barely started, and already it feels like a well-rehearsed script. That's the moment — before everyone starts playing their default roles — where something different becomes possible. It's noticing when the pattern is emerging, and making a different choice.
These patterns hold not because of bad intentions, but because everyone is participating in a way that feels familiar — and even functional, up to a point.
Can it be named without triggering the next move? Can the pattern be interrupted in a way that opens a new pathway rather than just disrupting? This might require saying what's actually happening in the room rather than routing around it.
This level of honesty might feel more exposed. But it's the only move that operates on the variable that directly affects what's possible. The capacity to do this has less to do with personality than with practice — developing the specific skills for redirecting what's happening without escalating it, making room for what's not being said, and shaping the context of a conversation rather than just its content.
That's what relational skill actually is. Not just "better communication" or emotional intelligence in some vague sense. Not being a "people person" or extroverted. It's the ability to recognize a pattern and change your participation in it in real time.
Most of us were never taught to notice the water we swim in.
Turns out I wasn't the only one solving the wrong problem.
Which brings the opening question back around — what if technical expertise really is becoming less relevant?
As throughput accelerates, misalignment becomes more consequential. A team that can't surface tension early, repair it cleanly, or hold genuine disagreement without spiraling will hit that ceiling sooner, and harder. Coordination quality becomes the binding constraint, and it gets there faster than it used to.
Here's what makes this stranger than it sounds. The administrative layer of coordination — the scheduling, the tracking, the project management overhead — is exactly what AI is compressing. And yet people are talking more, not less. When the scaffolding falls away, what remains is the raw relational work. The project management systems were buffering human contact. Remove the buffer, and certain kinds of contact actually intensify. When the AI handles the code changes, the people have room for the conversations that actually matter, like what would we build if the technical details were trivial and why?.
The people stuff doesn't disappear — it takes center stage. As recent ACM research documents, AI is making software engineering more human centric.
AI is making software engineering more human centric.
Engineers who make the jump to management must learn to dial back rigorous attention to detail — left unchecked, it makes you a micromanager. AI coding agents have something of the same non-deterministic quality. Dialing back the need to control turns out to work for human coordination and coding agents alike.
If you think AI will eliminate jobs altogether — consider that world for a moment. The transition itself would be a human coordination problem of historic scale. In what comes after, navigating human relationships won't become less relevant. If anything, it would become the thing that remains (one hopes).
Relational skill doesn't decay when the technical ecosystem shifts. It isn't tied to a stack or a methodology. It compounds across teams, roles, and contexts. And unlike technical execution, it's harder to replicate or compress — because the constraint isn't feature throughput. It's the conditions under which human beings can actually work together.
That's what makes it worth developing deliberately. Not because the world needs another argument for soft skills. But because the specific capacity to see relational patterns as patterns and change your participation in them is both scarce and durable.
What helped me was having a framework for it — something that made the invisible visible. MetaRelating is built around exactly that orientation — the systems intelligence that makes this kind of participation possible. It's more than just communication techniques, but a framework for seeing what's actually happening in a relational system and developing the skills to move inside it differently.
The good news is that this is learnable — not as a personality change, but as a practical communication skill with identifiable components and real feedback loops. The same way technical competence was built, piece by piece, over time.
Most people in technical environments just never had a reason to start. The environment is changing that.
