Learning and Overlearning

Learn from your mistakes. Don't let the scars become identity.

This past fall, I was speaking with a founder friend about their company. I gave advice about how I’d approach data architecture were I running their startup. I was really adamant that I had the right answer; we had just done something similar with my previous startup, Atlas Urban Farms. There was a lot that went wrong with Atlas, but our data architecture was excellent. My friend ended up going with their gut, instead of my firsthand experience, and it really pissed me off.

As I reflected afterwards, I zeroed in on something odd. Why had I felt so strongly about doing things my way? I realized that I’d tried applying the lessons from Atlas, in the hopes of preventing another startup from going under. My mistake was assuming my experience applied universally. There was also the fear that if my lessons couldn’t be applied elsewhere, then their validity might be undermined. The time and effort I spent gaining this information might have been wasted. This really threw me for a loop, undermining all the reflection I’d been doing. Could it be possible not just to learn lessons from my recent startup, but to overlearn them?

My thoughts turned to history, one of my most common and trusted touchstones. “Generals always learn to fight the last war, especially if they won it.” Georges Clemenceau, French Prime Minister during World War I, gets credit for that quote. It’s easy to find examples of this throughout history. In 2026, we might think it insane to run directly into machine gun fire, but to the generals of the First World War, that was the winning playbook. They were simply applying tactics they knew would win, based on the wars of the 19th Century. Why fix what isn’t broken?

There’s more danger learning from wars lost than wars won. It’s easy and comfortable to continue what’s been working. You have a track record of success, and change is hard. The threat is the inertia of success lulling you into stagnation. From a loss, the threat is different. Scars become lessons, calcifying into identity. What might have been tactically flexible in the past becomes scripture, unquestionable because invalidating those tactics would lay bare your own shortcomings and failures. The more we cling to those lessons, the more rigid we become, and the less able we are to face future challenges on their own terms. Another illustration from history:

After WWI, France concluded that a stronger defense would win the next war against Germany. The French army’s rallying cry from the Battle of Verdun, “They Shall Not Pass,” became a national ethos. The French built massive fortifications, but became stuck in a defensive mindset. Meanwhile, the Nazis had developed a more mobile and technologically advanced form of warfare known as blitzkrieg. They went around the fortresses and routed both the French and their British allies. French defenses like the Maginot Line would have dominated the First World War, but the world had moved on. The French were fighting in 1918, and the Germans were fighting in 1940. The mistake cost France its independence and extended World War II by years.

I believe in never making the same mistake twice. I went searching through my mind for Lessons I could pull from the experience in the early days after my decision to shut down Atlas. In the perfect world that existed in my head, I wanted to produce a list of rules informed by my failure to make Atlas financially viable. Maybe it would be a Top Ten List; How to Build a Business, with real life examples. They would be like laws of physics for the next business I wanted to start, and also for the countless founders I mentor and advise. Part of this was to prove to myself that Atlas was not a total failure. Here was some value that might make the past four years of effort worth more, to me and to others. That was my intention, bringing my experience to the conversation that opened this piece.

The challenge is internalizing these new lessons while not letting them eclipse older knowledge. In the process of reflecting on this, I thought that I was writing about confirmation and recency biases. Confirmation bias is the phenomenon wherein you seek out or more readily accept information that confirms your prior assumptions. Recency bias is the tendency to overemphasize recent information versus older information. Upon further reflection, I realized I was using these concepts as a shield. I’ve really been writing about vulnerability, and the idea that we make mistakes, and that the solution to those mistakes is not always something we can apply to future problems. It’s a much more personal, identity-based lesson, and one that did not become clear to me until I had written and rewritten this piece a few times.

I now know to be on alert if I advocate for anything too strongly after the data architecture talk and reflecting through my writing. It’s a process that I’ll keep learning, trying to improve each day. I’ve also learned to be wary of things that can threaten my core identity as a founder and a problem solver; not that it would be possible to eliminate these triggers entirely, but rather to know when I need to invest a little bit more time and energy into keeping myself balanced. It’s a constant process. There’s no real end to it, no point at which you just get to coast and things are perfect. I can’t offer you a perfect solution, but I can give you this advice:

Learn from your mistakes. Allow for new paradigms. Don’t let them become dogma. Past lessons should guide rather than constrain. It’s not easy, and you’ll always be working with a shifting model. Stay mindful, and good luck.

Photo by Provincial Archives of Alberta on Unsplash