There’s a fascinating debate swirling around the impact of Artificial Intelligence on worker productivity, sparked by thinkers like Ethan Mollick. In his book Co-Intelligence, he suggests AI will be a great “leveler,” disproportionately helping lower-performing workers catch up to their more productive peers. But is that what’s really happening? After years of using AI to accelerate my own work, my experience points to a different, more nuanced reality.
The idea that AI primarily benefits those who are struggling, while top performers have less to gain, doesn’t quite hold up.
“My personal experience over the last year does not match that assessment. What I’ve observed is that the ability to effectively leverage AI benefits maybe not those who are most productive immediately, but it does not have some sort of bias to serve and benefit people purely based on a productivity spectrum.”
Instead of being a simple leveler, AI appears to select for a specific, multifaceted set of skills. Excelling in this new era isn’t about where you start on the productivity scale; it’s about your ability to cultivate the crucial skills that allow you to effectively wield these powerful new tools. Here are the five skills that truly matter.
1. Humility
The first, and perhaps most critical, skill is humility. As an engineer, you have to accept that some of the work that once brought joy and a sense of accomplishment is now better handled by an AI. It requires staring at your own potential obsolescence in certain areas and having the optimistic perseverance to adapt. Without this humility, it’s easy to fall into the trap of discounting the magnitude of this shift, claiming “this is not a Moore’s law type of thing,” and explaining away the revolution happening right in front of us.
2. Architectural Sense
A strong architectural sense is paramount. Knowing where to draw boundaries, how to define levels of abstraction, and which way dependencies should point is more important than ever. You need to understand the fundamental structure—the “dag of my software”—to guide AI effectively. This isn’t just about writing code; it’s about designing the blueprint that the AI will build upon.
3. Ergonomics
This might sound strange, as ergonomics is typically a human-centric concern. But it turns out that AI, like people, performs better in a clean, well-organized environment.
“AI is great at managing complexity, but it’s better at managing simplicity. If an AI comes into spaghetti code in an arcane language, it’s going to be less productive than if it steps into something that is well-designed, well-architected, and well-documented. Just like a human.”
Creating simple, ergonomic systems with low complexity allows an AI to understand the context and get to work faster, maximizing its productivity.
4. Deep Systems Knowledge
To direct an AI, you must first understand the landscape yourself. This means having deep, intuitive knowledge of how large-scale software systems are built. You need to know your tools, whether it’s AWS or other platforms, and understand the trade-offs between different frameworks, languages, and infrastructure. This foundational knowledge is what allows you to make smart, high-level decisions and guide the AI with an expert hand.
5. Clear Communication
Finally, once you have the knowledge, you need the ability to communicate it. This isn’t about writing longer prompts; it’s about speaking with clarity, precision, and confidence. It’s about defining concrete success criteria and providing guardrails, much like a great manager empowering their team. You must be able to tell the AI exactly what you want, what success looks like, and what boundaries it must operate within. This skill directly impacts the quality of the output.
So, Is AI a Kingmaker? 👑
When we zoom back out, the question remains: is AI a leveler or a kingmaker? I believe it is a kingmaker. But it’s not necessarily crowning the engineers who were already the most productive. The old measures of productivity—like lines of code written—are becoming obsolete.
Instead, AI is making kings of those who master this new blend of skills. The individuals who combine humility, architectural vision, and crystal-clear communication will be the ones who thrive. The most productive engineer of yesterday might lack the humility to set down their keyboard and adapt, proving that in the age of AI, the crown doesn’t go to the fastest coder, but to the wisest director.