Stop forgetting the plan

One of the biggest issues I run into with AI agents is that the implementation of a given feature starts to drift from the original implementation plan.

This is extremely frustrating, especially after spending a significant amount of time iterating on a plan.

And frequently, these misses are subtle and hard to catch.

The reason this seems to happen is because as the context fills up, the specific implementation details in the plan itself get lost. Basically…the agent starts to forget parts of the plan itself.

An interesting observation is that the now forgotten “Ralph loop” approach actually addressed this point well. Because it broke tasks into small granular and isolated units of work and then addressed those small tasks 1 by 1 each in a brand new chat, it kept the agent aligned and grounded.

Plan mode in Claude Code and OpenCode doesn’t quite work this way.

So what is the fix?

When the agent generates the Task list or Todos, I ask that every 3 tasks, it inject a task to re-read the entire plan file.

This means the plan keeps getting freshly injected into the context, so even during a long 1hr+ run, the agent remains grounded and aligned to the plan.

The time to build is now.

There has never been a better time to build those big ideas, the ones that required funding and entire teams and months or years of effort to deliver.

AI has changed the game and it’s made it easier than ever for the little guys to compete with the big players.

There are so many ideas that I have been sitting on because they were simply too big for one person to take on alone. Now a single $200/mo subscription is the equivalent of a full supporting team.

I have been a software engineer for a very long time. I have worked on over 30 different projects spanning every single part of the stack and all sorts of paradigms: monoliths, microservices, mobile apps, web apps, APIs, content sites, data pipelines….you name it.

The point is, I have experience, a lot of experience. And I can say without a shadow of a doubt that these tools work.

BUT…. you have to know how to use them effectively. You have to hone your skills, you have to practice, you have to really put in the work. I mean…that’s the nature of software engineering as a discipline..isn’t it?

You have to become a proficient operator of these tools.

The best way to do that is to BUILD. Build something real. An actual product. Not a toy app you will never ship.

Push yourself.

Go through your notes where you jot down all your ideas, find the one that you were most afraid to take on and start and don’t stop until you ship.

The time to build is now.