I Tamed AI’s Verbosity with a 50-Year-Old Standard. Here’s How.

You know that feeling when you read an AI-generated document and you want to scream? The endless run-on sentences, the vague warnings, the passive voice that sounds like it was written by a committee of lawyers having a stroke. You’re not alone. Every engineer, technical writer, and ops manager has wasted hours untangling bloated AI output. We’ve been told to ‘just prompt better’ or ‘add more context.’ But that’s like telling a drowning person to swim harder.

AI doesn’t have to be a bloat machine. It can be the most disciplined editor you’ve ever hired.

I discovered a project on GitHub that does exactly that. It’s called SimpleEnglish, and it forces AI agents to write in ASD-STE100 — Simplified Technical English. This is a 50-year-old standard originally developed for aircraft maintenance manuals. It’s the opposite of ChatGPT’s default style: short sentences, active voice, no ambiguity.

One developer tested it and shared the output:

“Before start, ensure AWS credentials are correct. Otherwise, S3 rejects uploads with a permission error.”

Compare that to the typical AI output: “Before you begin the process, it is important to verify that your AWS credentials are properly configured. In the event that they are not, the S3 service will reject the upload attempt and return a permission error.”

Which one would you rather read in a crisis? Exactly.

Here’s the mind-bending paradox: we’re using billions of parameters — the most complex neural networks ever built — to output the most rigidly constrained, simple language possible. Extreme computational complexity achieving extreme linguistic simplicity. It’s like using a supercomputer to write a haiku.

But that’s the point. The problem with AI-generated text isn’t a lack of intelligence. It’s a lack of constraint. AI models are trained to be endlessly helpful, which means they never shut up. They hedge, they qualify, they add fluff. And that fluff kills clarity.

Everyone is trying to make AI smarter. The real breakthrough is making it dumber — in a controlled, precise way.

This isn’t just a hack for documentation. I’ve seen people in the comments say they want to use it for emails and Slack messages. Because clarity is a superpower in any communication. The standard forces you to drop the jargon, the passive voice, the weasel words. It’s like having a ruthless editor standing over your shoulder.

Of course, some people dismiss it as “this week’s mindless productivity trend.” They’re missing the point. This isn’t a trend. It’s a response to a fundamental failure of modern AI: the assumption that more output equals better output. It doesn’t. Better output is shorter, clearer, and more actionable.

I’ve started using it for my own technical documentation. The difference is night and day. My team no longer asks “What does this mean?” They just execute. And that’s the ultimate test: If your AI-generated text can be misinterpreted, it’s not intelligent. It’s noise.

The next time you read a bloated AI document, remember: the solution isn’t more AI. It’s less. Less ambiguity. Less verbosity. Less noise. And that’s the most intelligent thing we can do.