2026-05-20 - "AI Rewards Makers", or: EM -> IC
May 20, 2026
In the last 6-12 months, many friends and former coworkers have asked me: “Allen – I saw you switched from management back into IC. What is changing about Engineering Management, and what should I change about my career?”
What’s Happening
I had the same question for myself about a year ago. I’d been managing for 8 years. I was increasingly frustrated and exhausted. My stress levels were at an all-time high. When I reflected, I came up with three major reasons:
- AI let you “just do things” and output the code of 10 people. I wasn’t “doing things”.
- I got into engineering management to be empowered to get things done and ship things.
Now, I felt increasingly less “blocked” (i.e. needing to organize people) to ship things. The speed of getting things done was limited by the speed of my own ability to clarify my thoughts.
- I got into engineering management to be empowered to get things done and ship things.
- My job was increasingly about managing personalities and people, and less about the “maker” parts of management: gathering people together and working through something together like figuring out the gnarly tradeoffs of a new API, or the complexity of shipping a new product.
- The work shifted toward delivering tough feedback, performance management, firefighting org issues, mediating disagreements. These are real and important, and I don't want to suggest otherwise, but it’s difficult to balance this concurrently with the creative “maker” job, and the ratio kept tilting.
This drained me too much to actually have “maker energy”. Even though I had more hours in the day, I didn’t have energy left in the day.
- The work shifted toward delivering tough feedback, performance management, firefighting org issues, mediating disagreements. These are real and important, and I don't want to suggest otherwise, but it’s difficult to balance this concurrently with the creative “maker” job, and the ratio kept tilting.
- I felt strongly that there wasn’t a path to responsibly adjust my management responsibilities to enable me to do more building.
AI Rewards Makers
I believe AI rewards makers. I’m seeing the industry accelerating a shift to an insatiable appetite for “makers”; and a shrinking appetite for “managers”. I see this most clearly in hiring patterns. I believe longevity for management careers means preserving the “maker” skill.
“Maker” is orthogonal to IC or EM. It is purely a function of where energy is spent. We typically only associate ICs with “makers”. But one can also be a VP of Engineering who is deep in the architecture and code with their team, working through the implementation details with the team. That is a maker, regardless of their formal title.
Based on my own experience, my own advice for those who enjoy the “maker” job is:
- Ensure your job lets you be a maker. In most management jobs, you get to define your own role. Just make sure it doesn’t saddle you with so much overhead that you can’t create something.
- If your job doesn’t let you do this, consider becoming an IC. That will be a simpler, more natural forcing-function to align your job description with the “maker” job.
Making the Leap
In pursuit of developing and preserving that “maker” skill, I began planning and lightly testing an approach to see if it’s compatible with my current role.
Be relentless in minimizing scheduled “manager” work. Manager work refers to “the things around the thing”. Usually it’s scheduled meetings, and things that break up your day. They are necessary, but often reduce capacity for creative work.
Being effective with less in a management role has always been a critical component of the job, but the pace of execution now even further accelerates the need to find leverage in the role.
Reflecting on my personal experience, I found this meant:
- Increasing protectiveness of energy. For me, time wasn’t the issue: energy and focus was. I began mapping out types of work that drained me, and, on designated days, avoided them through calendar tetris. I experimented with the 1:1 format: less-than-weekly cadence, more informal catchups, mixed in with group formats. Some reports preferred this, others didn't, and I adjusted. I pushed back on the idea that weekly 1:1s are a uniform good, and adapt to individual styles. I found group panel meetings to be an exceptionally effective way to utilize 30 minutes.
- Making decisions more quickly. I was out to dinner with a good friend the other week, and we laughed about how we would agonize over reorg plans at Plaid for weeks. Many companies reorg every few months at this point. With the pace of the industry now, those look like obvious two-way doors in hindsight.
- Leaning harder into hiring for and fostering high-agency in your team. Shifting toward situational coaching over scheduled coaching. This isn't right for every team or every level (a team with more junior engineers needs something different) but for some teams, it worked.
- Preserving team health. I found this difficult, because it meant balancing being “present” versus maker energy. Many things I did to “minimize manager work” impacted team health. Less frequent 1:1’s, leaner coaching – this worked for some, especially the more senior engineers on my team, but it wasn’t free. I’m sure there were moments where individual reports needed more from me than they got. What made it even more challenging was that ~80% of my organization was not in San Francisco, making serendipitous connection more difficult, especially for new hires integrating into the team.
Do more of the “maker” work. Put simply, maker work is: decide what to do, then do it.
I found some pleasure in building things that simplified my management job – I played with building a personal performance review copilot, a personal financial planner, and a personal health tracker.
Some manager work can be reframed as maker work. In lieu of 6-month plans, we sought to prototype earlier. We expressed ideas through prototypes. I felt this was always a good idea, but AI has dramatically reduced the effort-to-prototype.
I also felt a visceral need to take a significant builder role in something on the team. Unfortunately, for reasons below, I wasn’t able to do this.
If you’ve tried this and it’s not working, consider being an IC. I switched to IC because, even after all of this, I still felt unsatisfied with the amount of maker energy in my day. I didn’t manage to get a foothold on contributing to projects directly on my team. Part of it was the workload, but another part of it was some personal burnout. I didn’t feel confident there was a reasonable path in the management role, where the maker work dominated the manager work on a reasonable timeframe. Waiting 6-12 months felt like an impossible task.
And perhaps most importantly, the cumulative toll of reduced 1:1’s, leaner coaching, and just a lower presence on the team was just not the right tradeoff to make for my team. The job was truly fundamentally incompatible with a “maker” role. The team deserved a structure that gave them the support they needed.
Hindsight
9 months later, I think I made the right choice. I get to focus all my energy on building. I get to be an active member in the industry transformation. Building with AI has weirdly been this positive-feedback loop that gives me more energy and focus to do the next thing. Also, I feel I’ve gained a deeper understanding of my strengths and gaps as a manager. If I ever feel like that’s something I want to do again, this experience is one of the most trajectory-changing things I’ve done in leveling up as a manager.