I'm 71. ADHD and dyslexic. I've been building enterprise systems for 45 years — mainframes to web apps.

Two years ago I built a CBT-I sleep app. It failed.

The failure wasn’t technical. It was cognitive. I designed a tool that assumed users had full executive function available when they used it. They didn’t. When people are depleted, working memory narrows, inhibition weakens, emotional reactivity increases, and sequencing suffers — exactly when structured tools demand those abilities.

That got me thinking: most productivity tools and corporate training assume optimal cognitive state.

But professionals rarely operate in optimal state all day.

So I built a simple model around four capacity states (Green, Yellow, Red, Shutdown). The core claim is:

An hour of cognitively demanding work in a depleted state produces less net value than an hour in a resourced state — and in some cases introduces negative value through rework or error.

To make that concrete, I built a calculator that estimates annualized loss based on:

* Salary (or value of time) * Estimated hours spent in each state * Estimated efficiency degradation %

The default degradation values (20% / 35% / 60%) are derived loosely from cognitive load and sleep restriction literature, but they’re obviously simplified.

What I’m looking for:

1. Are the degradation assumptions directionally sane? 2. Where would this model break down for engineering work? 3. Is there better literature on state-dependent productivity loss that I should be using?

I’m not looking for signups. I’m looking for holes in the model.