Is Engineering Management Obsolete?
AI is flattening the org chart. I think that's a mistake.
No frameworks or playbooks today. I just want to put out my thoughts on the role of engineering management in the age of AI.
When you read LinkedIn posts and industry news, you are bound to stumble on claims that everyone is a manager now, managing a bunch of agents. How organizations are flattening. How AI makes it possible for one manager to supervise teams of fifty engineers. That middle management is not needed anymore.
These are strong opinions from industry leaders. They cannot be easily dismissed. Or?
If I learned anything during the last decade, it is that nobody knows anything for sure. All the political turmoil in the world, COVID, geopolitical conflicts, LLM-based intelligence (aka AI-revolution) were black swans. Something that the public didn’t see until it happened. When they happen, people rationalize, and experts write books and articles about how it was all inevitable.
So, I will not pretend I know the future. I am just trying to think from the first principles and trying to recognize patterns from the past, whatever it is worth.
I’m Roman, CTO at a startup, and I write High-Impact Engineering — a weekly newsletter for engineering managers, directors, and CTOs.
Below is a short opinion piece on the future of engineering management.
The change is the only constant
The nature of work and what is hot on the job market is constantly changing.
Thirty years ago, when I was in elementary school, my grandfather was saying, “Roma, you should go and study how to fix TV sets, you won’t ever be without a job.” We all know how this advice aged.
Now my own teenage kid asks me what he should study to be in demand in the job market. He is thirteen years old now. It is five to twelve years before he enters the workforce. I don’t know what will be hot on the job market so far into the future. I told him so. I don’t know which technical skills will be in demand. What I am pretty sure about is that leadership and management are timeless.
Fundamentals stay
As long as people need to work in groups to achieve common goals, there will always be a need for someone to lead, inspire, and take responsibility on behalf of others. It has been true throughout the whole history of human civilization. Tools have changed from stone to bronze, to iron, to machines and computers. The fundamentals of working in groups haven’t changed.
This is why I believe that engineering management is not going anywhere. If anything, it gets more important in the age of AI.
When execution gets faster, the individual leverage gets bigger. It means that both the positive and negative impacts of individual contributors and groups have increased.
I believe that experiments that Facebook, Block, and others are running now will be short-lived. The logic behind them is that managers mostly route information up and down the hierarchy, and AI can now do that routing. I don’t think it has ever been the job. Routing information is an important but small part of management. Aligning people, resolving conflict, maintaining business and technical contexts, growing them, inspiring, and taking responsibility on behalf of others is the actual work, and none of that is something you can delete and hand to a model. I feel bad for people in these organizations being guinea pigs, but I think in the long run it is interesting to see what the outcomes of these experiments will be and whether we learn anything useful from them. I am personally skeptical, but learning is learning, even if I consider these experiments reckless.
What does it all mean for me?
Talking about reckless, I think it would be reckless of me to finish this post without any practical takeaway.
Here it is. The best investment for a manager right now is not to learn new commands in Claude Code, build a prompt library, or jump both feet into agentic coding.
It is looking at the bottlenecks in the current process, thinking of what has become possible, and whether old best practices are still the best way to go. Look at the team and organizational health, invest in psychological safety and effective communication. Identify your own shortcomings that affect the team’s effectiveness and well-being.
These things are timeless. Tools change, but “basics” proved themselves to be something that endures through generations of leadership, and I dare to say millennia.
I love the Grandfather story. We have shared nostalgia! Mine was a radio engineer who pushed me towards computers.
Agree with you on EMs remaining important. As you say, developing those core skills - communication, conflict resolution, etc. - is more important than ever.
Facebook and other large orgs can withstand failed experiments and easily restructure, smaller businesses cannot.
The information routing point is the right reframe. The people calling management obsolete are confusing one function of the role with the whole role. What they're really saying is AI replaces the status-update layer. Fine. The alignment, conflict resolution, and accountability layers are getting harder, not easier, as execution speeds up.