Getting Uncomfortable Is the Only Moat Left

Every job I’ve had has ended up the same way: find the thing that’s broken, model the case for fixing it, build a prototype and walk it to production.

AI makes the building part cheaper every month. It does not make the walking part any easier.

The walking part is persuasion, trust and the willingness to be wrong in public. Those are learned by doing uncomfortable things on purpose.

So I’m doing a few. I’m putting my face on a short-form page. I’m starting a consulting practice instead of waiting for a company to promise long-term security. I’m planning a neighborhood game night with people I haven’t met yet.

None of these require a model. All of them require a person willing to feel awkward for a while. That, I think, is the durable skill.

It is tempting to treat discomfort as a phase, something to get through before the real work starts. I have come to think it is the real work. The moments that felt most awkward in past jobs were usually the ones that moved things forward.

Most of those moments were conversations: telling a leader a project was off track, asking a customer why they left, admitting a prototype did not do what I had promised. None of them got easier with better tools.

That is the part I want to practice in public. Not because anyone needs to watch, but because doing it where people can see raises the stakes just enough to make the practice count.

If it works, I will have more of the skill that AI cannot hand me. If it does not, I will have a record of trying, which is its own kind of useful.