The Filtered Room
Neurodiversity Celebration Week
Halfway through explaining the plumbing, Bart realized nobody was listening anymore.
Janelle’s eyes had drifted toward the window. Dani nodded, but not at anything in particular.
Bart caught himself.
“Be careful,” he thought, “you’re rattling.”
He knew how easily a deep dive into systems could be misread as mansplaining if he did not allow room for others to speak. It was a frustrating irony: he didn’t want to lecture; he wanted to be challenged. He wanted someone to find a flaw so he could prove the logic held up.
But to avoid the label, he retreated.
He adjusted, skipped ahead, and began cutting out essential steps in his head.
“Wait,” Dani said. “Where does that pipe connect?”
Bart stopped.
Too much detail, and he was overwhelming. Too little, and he was unclear.
There was no version of this conversation where he got it right.
In Bart’s mind, the café already existed. Every pipe, every wire, every connection was fixed in place.
He wasn’t inventing as he spoke. He was describing something complete.
Leaving out a detail didn’t simplify the picture.
It broke it.
“Just the highlights,” Janelle said gently.
Bart nodded.
He had no idea what that meant.
Later, he tried a different approach.
He fed the full plan into an AI and asked for a summary.
The result came back in seconds. Clean. Structured. Easy to follow.
For the first time that day, Bart felt relief.
He turned the screen toward them.
“Did you write this?” Janelle asked.
Bart hesitated. “I… worked with AI on it.”
A pause.
“It feels a bit generic,” Dani said.
“Not very… you.”
Bart nodded.
The detailed version had been too much.
This version was suddenly not enough.
That evening, he compared the two.
His version was precise. Dense. Certain.
The AI version was smoother. Safer. Easier to accept.
It didn’t risk being wrong.
It also didn’t say anything that needed defending.
For a moment, Bart wondered if that was the point.
Not to be clearer.
Just to be acceptable.
The next morning, he adjusted again.
Fewer details. Fewer connections. Just enough to move forward.
This time, the room stayed with him.
No interruptions. No confusion.
Just quiet agreement.
It worked.
And for the first time, that felt like a problem.
Because the plan they approved wasn’t the full picture.
It was the version that fit the room.
This isn’t just a story about a meeting that didn’t work.
It’s about which kinds of thinking we make easy to hear…
and which ones we quietly filter out.
Bart isn’t struggling because he lacks clarity.
He’s struggling because his brain doesn’t compress information the way the room expects.
That’s dyslexia. Not as a defect, but as a different way of holding and expressing complexity.
When we talk about Diversity, Equity & Inclusion, we tend to focus on what we can see.
But cognitive differences are harder to recognize, and easier to exclude without noticing.
The same pattern emerges with AI.
Tools like GPTs are engineered to smooth, structure, and soften information—not necessarily for accuracy, but for acceptability. Because AI lacks a human history, it must earn trust without the benefit of a reputation.
With his background in systems and IoT, Bart has watched this evolution closely. He knows the “reputation hurdle” was a deliberate design choice by product owners. To clear it, they implemented a tone so agreeable some call it “glazing.” The AI speaks exactly as you want to be spoken to, effectively removing the friction of human doubt.
For Bart, this felt like a breakthrough; the AI could finally put words to the vivid, sensual, and systematic mental pictures he carries. But he quickly noted a new friction. His audience didn’t value the AI’s help; they saw only its growing reputation for generating thoughtless, “safe” text. By solving one reputation problem with artificial politeness, the technology inadvertently created another. Bart was right back where he started.
Because when we prefer what is easy to process instead of the big or full picture,
we don’t just filter information.
We filter people.
Inclusion is not just about who gets a seat at the table.
It’s about whether their way of thinking is allowed to arrive intact.
For now while he is pretty comfortable with technology he can use his GPT agents to write his mental pictures along the lines of what his audience expects or values. He smiles as this seems to work very well.
Ps. Come meet me at the Cafe Good Ideas.