Ambition and AI

Over the last year or so, I’ve had a lot of fun watching people get the “Claude Code high.” If you’ve seen it or experienced it, you already know what I’m talking about. If you haven’t, allow me to explain. There is this moment when a person realizes that they have this new “power” to create something that was previously completely inaccessible to them. For example, they can turn an idea for an application into a working version of it in a weekend. The symptoms include euphoria, excitement, and a lack of sleep due to working absurdly long hours with this newfound power. Look, as someone who runs an engineering org, I’ll be the first to admit that there can be some sticky situations when someone starts using a coding agent for the first time. But this isn’t about that (that would be a long, semi-ranty post where you’d like me less after you read it). Instead, I just wanted to take a moment and appreciate how incredible this is.

There are a lot of valid concerns around LLMs and the agents they power. Training data violating IP laws, environmental concerns, job loss, rapid adoption smoothing out our brains—the potential issues are legion. I’m not dismissing the above or making the argument that AI is worth it. What I will say is this: it’s pretty incredible what is now within reach for so many people. What really resonates with me is seeing people trying to make their ideas happen. There’s something deeply hopeful about that. I’m trying to get better at that myself: getting past my own pessimism and fixed mindset to just get out there and run a bunch of experiments. This is unknown territory, and I want to see what’s possible. Felix Rieseberg has a whole post on how to do this.

I think a lot about loss of skill and how to keep adding value as a human (and an engineer) in this “new world,” but today I’m thinking about how agents have created an opportunity for us to express our own agency—turning thoughts and ideas into reality—more powerfully than ever before.