The Manager–IC Pendulum and the Rise of the “Builder with Taste”

These are my personal thoughts, experiences, and opinions, and they do not reflect the views of the company I work for.

I’ve been reflecting a lot lately about the pendulum swing between Engineering Manager and Individual Contributor. For those who have not read it, I would suggest reading Charity Majors’ multiple posts about this topic and coming back here. It’s an emotional space I’ve been in for the last five years as a hands-on manager, but the last twelve months have completely changed the physics of that swing.

I’ve been keeping up with AI chat interfaces and Multimodal prompts since early 2023 when they started becoming prominent. I started using AI coding agents back in May 2025, and honestly, I was blown away. Since then, with the arrival of newer foundational models of Claude Opus and GPT Codex, the frontier of what’s possible hasn’t just moved — it’s teleported at warp speed.

I’m currently heading into a phase of my career defined by leading a team of highly motivated people working on a common goal, the nuances of middle management and providing architectural leadership. I love that work. I like solving problems along with a bunch of great people, understanding the business, and driving outcomes. But recently, I’ve been going on weekend binges building side projects and experimenting with AI-assisted coding. I migrated this blog to Markdown, built on new ideas for hobby projects, and watched that code spring to life on my GitHub repos and in my Vercel space, all co-authored with AI. This was an attempt to see how I can bring these ideas back to my day-to-day leadership, pairing more effectively with senior engineers, asking better architectural questions, and creating a space for experimentation in a responsible way.

Through that journey, it has become very clear to me that we are moving, or have already moved toward a world where the vast majority of code is AI-assisted or in many cases initially AI-generated. People worry about AI taking jobs. I see it differently. I think it removes the constraints of syntax and typing, but it doesn’t remove the craft. The bottleneck has shifted. It used to be how you build and how long it may take. Now, the bottleneck is what you build and how long you may spend re-doing it.

It reminds me of the shifts that happened in the music industry. Years ago, you needed a professional studio to record a song. Then came the Mac/PC and a digital microphone to have your own Home Studio. Suddenly, anyone could record a track in their bedroom. Did that make everyone a pop star? No. We still have the Ed Sheerans and Justin Biebers who have that X-factor compared to many others who had access to the same YouTube they first started with. But it did mean that building a brand, a reputation, and “taste” became more important than access to a recording console.

We are seeing the same thing in software engineering. We’re going to have a lot of senior engineers who can architect and build meaningful systems end-to-end because they have high agency and can think about systems. But it also makes me wonder about the rest. This keeps me up at night: How do I create an environment where mentorship, protected learning spaces, and building strong technical foundations are now more important than they’ve ever been? In the old world, you had a safe test bed to fail, break things, and learn.

I recently had a moment that crystallized this for me. My wife, who is non-technical but has great product sense and is learning AI, was building an app on Replit with my guidance. She was typing away, prompting the AI agent, and seeing real results. It was magic—until we ran out of credits. We stared at the screen. The magic stopped. And then it hit us: You either pay for more credits, or you figure out how to code.

Even with the best tools, you eventually hit a wall where you need to know how the machine works. You can have the GarageBand setup, but knowing how EQ works, how to mix, and why something sounds bad—that is still the difference between noise and music.

Business needs haven’t changed (yet). They want to move faster and deliver more. The classic triad “Good, Fast, Cheap” — pick two — is being stressed. I’m realizing that the value I bring isn’t just in prompting code. It’s in understanding what happens when that code scales. It’s easy to build a prototype for 1K users with Claude. But what happens when you hit 10K DAU or even 1M DAU? What happens when you have petabytes of data and distributed system failures? At scale, in large organizations with real customers, these tradeoffs matter more—not less.

Those core engineering problems aren’t going away. The rewrites and the refactors if things are not “built right the first time” are where the real cost is going to be. And this has always been the case, except your favorite AI model can vet your design before a single line of code is written. Your own personal Software Architect can verify your thoughts, but can you trust it enough to solve all of your technical tradeoffs? I do not know, but these types of questions spark an excitement in me about building things again.

So, am I a Manager or an IC?

I think those labels are dissolving. We are heading toward a consolidated title. Maybe towards something like a “Product Architect Engineer with Taste.” This is not the literal title but more of a description of how the roles are converging. The lines between Product, Engineering, and Design are blurring into a single role of “Creator.”

It is an interesting time to be going through such a transformation. Change feels like it is here now, not some distant future. If you feel the same way and are out there right now upskilling, using these agents to build faster than you can imagine, I’d love to hear from you on how you are bringing value back to your teams. It might not be perfect code you are churning out, but as the saying goes, don’t let perfect be the enemy of good. The future I see is in building systems that let organizations trust what AI generates while remaining secure, observable, and resilient at enterprise scale.

I don’t see this as a rejection of engineering or management, but as a convergence of the two. I’m of the opinion that the next generation of leaders won’t fit neatly into titles like Managers or ICs. They’ll be builders with taste—people who understand both architecture and empathy, who move fluidly between systems and storytelling. That’s where I want to keep swinging: somewhere in the middle, building with intent and a little bit of heart.