My mom said to me recently that one thing engineering school really gave me was the ability to teach myself.
It did not really strike me when she said it. It was just in passing, and I do not even remember what it was in reference to. When I reflected on the day later, it hit me that what she was describing was agency.
I have been trying to digest what high agency means in the new world and how people are going to be successful as the world changes. I keep hearing this term in a lot of the content I consume. I think it means an extremely self-motivated person who is willing to learn, teach themselves, and keep going without external motivation.
It can mean taking initiative, making an informed decision, and bringing other people along through what you do. I pair it with strong opinions, weakly held. You form a view and act on it, but you need to be willing to iterate and evolve as you continue to learn.
I Learned by Not Doing My Homework
I struggled to get my electrical engineering degree at Penn State. Most of the time I had to teach myself different concepts using YouTube videos or whatever else worked.
I also had a nontraditional approach to college. In some senior-level electrical-engineering classes, homework could be 10 percent of the grade and take 10, 12, or 15 hours a week. It did not always correlate directly to the exams, which were insanely difficult and graded on a curve to begin with.
To me, all of that work for 10 percent was not worth it. I was willing to sacrifice starting at 90 and work from there. That is definitely a controversial choice, but it is how I thought about valuing my time and how I wanted to spend it. I would often ask myself whether, when I looked back on college, I would be glad I labored over some electrical engineering homework. The answer was always, “No.”
My girlfriend at the time, now my wife, used to rag on me because I often did not take a backpack to class. A lot of my classes were death by PowerPoint. All the PowerPoints were distributed, so to me it was kind of a waste of time to take notes. I often did not have a backpack, or did not have much in it if I brought one.
Did that approach work? It worked well enough, depending on the answer. I had a 2.5 GPA. I also had a job locked up in October of my senior year, likely because I was honest about my GPA and the approach I took to school.
I spent a lot of that time with THON, leading a team of 37 for my last two years, and always made time for friends beyond that. Those choices did not turn the 2.5 into an academic success. They show what I was choosing to spend time on and the cost of doing that.
I was confident that my personal skills and hard work could make up for the lack of GPA. The part I see more clearly now is that I had confidence in my ability to teach myself and keep figuring things out, even though I did not understand that was agency at the time. I did not have a clue what agency was, to be honest.
Move Fast and Break Things
With AI, I basically just dive in head first. I watch a ridiculous number of YouTube videos, read blogs, listen to podcasts, and figure out how Codex, Claude, and different coding harnesses work. T3 is my current favorite. I am willing to push myself and learn outside of business hours, then figure out how to translate it back into the job.
There is a really fast feedback loop. You can actually see things quickly when you are building software, websites, spreadsheets, or prototypes. That feedback loop is like an adrenaline rush that keeps me going.
I enjoy sports betting and daily fantasy, and I think it plays into my personality. Even when things are hard or do not work, I have a relentless attitude about figuring them out. I can change something, see what happened, and try again.
The important part was realizing how fast I could get into a skill set I did not have inherently.
Knowing Enough to Be Dangerous
At work, back in 2024, we were programming a battery controller. I was using AI to help lay out the structure and think through how to develop an efficient program for a contractor we hired. These were the early days of ChatGPT and Gemini, but even then, the capabilities were mind-blowing to me.
That was a skill set I did not have. We hired somebody else because they had it. AI gave me the confidence to question the contractor a lot more: Why did you do it that way? Would this other approach work? I did not have to blindly trust work I could not understand.
I spent a lot of time fact-checking. I checked source material and used what I learned to ask better questions. AI still made mistakes, but it gave me a place to start and something specific to verify.
It changed my confidence in my ability to learn things extremely fast. I could get myself up to speed much, much quicker than I otherwise would have, or would have had the confidence to do.
That was a huge unlock: what else out there could I learn, figure out, and change in our business? It is a mindset I have carried since.
A Prototype Changed the Conversation
I had been beating the drum inside the organization that we needed to use AI, and it took years before the organization started to turn. The change really started when I began prototyping.
We were trying to win a large prospective customer. The request was not something our software could offer at the time, but it was not too distant from what the software already did.
I mocked up the entire setup in an interactive HTML file to demonstrate how it would work. We sent the files to the customer so they could play with it. The deal is not 100 percent done, but it got a notoriously difficult prospect very engaged with us. Even if we never close that deal, the experience and how it changed my approach will be worth it.
That prototype is what kicked everything off for how we do that kind of work today.
I have a version of the codebase in a local environment on my computer. I use different AI models to understand the codebase, the styling, and how the software actually works. Then I prototype a new feature: where it should be, how it should look, what it should reference, how it might connect to the database, and where it should pull information from.
The prototype is paired with a Markdown file explaining what I am trying to convey in the HTML. It is a visual paired with text for the developers to pick up.
Before, someone would make a rough mockup or an informal request of the development team. The developers were trying to paint somebody else's vision and had to assume what the person meant. Then there was a lot of back and forth.
The new handoff helped with all of it: speed, clarity, expectations, and execution. It enabled the developers to build things faster. They still own how the real feature gets integrated. I am giving them a much more concrete explanation of the thing in my head.
I also found a new passion for building things that I could not build before because I had access to new tools and learned how to harness them. I am an engineer, but I am not a software developer. Before these tools, a lot of these ideas would have stayed as a conversation or a rough explanation. Now I can make them real enough for somebody else to react to.
High Agency Compounds
I do not think high agency is a solo thing. I think you can lead a team with high agency, but I also think a lot of it is leading by example.
You are not necessarily going to talk somebody into it. If they see you doing things that are high agency, I think it can be contagious. I had been beating the drum about AI for years. The prototypes gave people something useful to look at, and that changed more than the talking did.
The point is not that I should win every argument. The contractor brought skills I did not have, so my responsibility was to question the work, fact-check my assumptions, and then listen to the answer.
It is the same with the developers. The prototype enabled them to build things faster, but I did not replace what they know. I showed the feature I had in my head. They are the ones who turn it into real software.
It takes one person to bring a different perspective or approach to the work. That example can inspire ideas from the rest of the team that one person would never reach alone.
Becoming the Most AI-Enabled Version of Myself
Bill Gurley has talked about becoming the most AI-enabled version of yourself. The interview that stuck with me is here.
I want to weave that idea together with what my mom said. The most AI-enabled version of me starts with the ability to teach myself.
I was willing to push myself and learn AI outside business hours, then figure out how to translate that back into the job. The battery project changed my confidence in my ability to learn things extremely fast. The prototype gave me a way to build things I could not build before.
I see that as the unlock: what else out there can I learn and figure out? That question changed my mindset.
I am not trying to sound smarter than I am. I hear “high agency” all the time and wanted to ground it in the examples that made it real to me.
My mom said engineering school taught me how to teach myself. Now I want to find out how far that habit can take me as I become more AI-enabled.