The obstacle is the way, revisited

When we build startups we're told we need to ship as fast as possible. We're often told to ignore the obstacles we encounter along the way, because they can distract us.

Only customers matter. We need to ship. Tests will wait. Documentation, CI, CD, linters, coding standards, all of that will wait. We need to validate our hypotheses. Most of our code will be thrown away anyway.

I think there is a lot of merit to this approach. I can't yet explain why I feel it's different for Bluewind.

I don't feel like what I am building needs to be validated. I am not really building a user experience. I am building the platform that will allow me to build any user experience.

So the obstacles I am facing while trying to build my user experience (a customer platform) are my absolute priority.

We'll see where this mindset leads me. I am not pretending I am right.

In the past I always had the "just ship it" mindset. I still think I have it. It's just that I have faced the same problems over and over again. So I know that when something annoys me about how the project is structured, it will annoy me at least 5 times a day.

And I see Bluewind as a 10 years project. Maybe more. The downside of thinking this way is that I make decisions that are very long term focused.

If it takes me 4 hours to add a command pannel that allows me to switch between actions in bluewind, I will do it.

If it takes me 1.5 days to build a something that automatically generate tests from the flows I created, I will do it.

If a data model is the correct one, but harder to work with, I will use it.

If it takes time to copy paste context from many files into chatgpt, I will build a context extractor.

This metawork is dangerous. It's like people working 100% on improving their productivity and forget about getting things done.

When you stop doing the work, the metawork becomes less relevant.

But this is also a superpower: most of the metawork engineers do ends up becoming a dev tool they sell.

But what happens when you use dev tool instead of selling it? What happens when you do the work AND the metawork?

I think what happens is that you end up building a laser focused tool. Because you don't have 100 of users telling you to shape this tool for their needs.

You just have it for yourself. This drastically reduces the scope of your work, but doesn't remove any of YOUR capabilities.