The Best-Disguised Procrastination Has a Roadmap
I almost spent a month building the perfect brain for my AI agents. The deals that pay rent were sitting unclosed the whole time. The bigger the build, the better it hides.
I have a note already about building a factory to avoid fifteen minutes of work. That one was easy to catch. The factory was obviously too big for four small tasks.
This is the grown-up version. The one I almost did not catch.
This week I started designing a single unified memory system for all my agents. One brain that knows my work, my clients, my conversations, my whole operation. Every session reads from it before acting. No more priority drift. No more re-explaining context. A real platform.
It is a genuinely good idea. That is exactly what makes it dangerous.
A great build can still be the wrong field
I had to say this out loud before I could see it. The unified brain is a multi-month build. And the actual constraint on the business right now is not memory. It is runway. It is closing the warm deals already sitting in the pipe.
So picture a farm. The crop in the near field is ripe. It is ready to sell this week. And instead of harvesting it, I am at the far end of the property drawing up plans for a beautiful irrigation system that will serve a field I have not planted yet.
The irrigation system is real engineering. It will work. It would help future harvests. None of that changes the fact that the ripe crop rots while you build for the harvest you do not have yet.
The plans feel like leadership. They feel like the responsible long-term move. They feel like the thing a serious founder does.
That feeling is the disguise.
Why the big build hides better than the small one
The factory-for-four-tasks was caught fast because the math was absurd. A day of building to save an hour. Nobody can defend that for long.
But a platform that unifies your whole operation? That has a roadmap. It has phases. It has a slide where the lines go up. You can talk about it for an hour and sound like a visionary the entire time.
The more legitimate the build looks, the longer it can stand between you and the work you are actually avoiding. Scope is camouflage. A small dodge gets spotted. A grand one gets a kickoff meeting.
And the work it hides is almost always the uncomfortable work. Following up on a proposal that has gone quiet. Asking a warm prospect to actually decide. Finishing the retainer task that is just polish, no novelty, no dopamine. Building for the agents is clean and infinite. Closing a deal is messy and has a yes or no at the end of it.
We do not run toward the build because it is the smart move. We run toward it because it has no rejection in it.
The test that cuts through it
I did not kill the idea. I shelved it, which is different. The honest reframe was this: that grand unified spine is a product. If a client funds it, it is the work. If I fund it out of my own runway while pipeline sits open, it is the avoidance.
So now the question I ask before any internal build is simple. Is this the crop or is this the irrigation? Is someone paying for this, or am I paying for it with the hours I should be spending closing what is already ripe?
If the constraint is cash, the build that does not serve cash is a hobby with good production values.
I still want the brain. I will build it. But I will build it when it is funded work, not when it is the most interesting way to not make a phone call.