by Ed Thompson
In 2006 I had a traumatic brain injury in a car accident, and my brain started working differently. My speed of processing changed, my working memory was impacted. I noticed how things that used to be automatic took deliberate effort, and I had to work out, over time, what this meant for my work.
It took me years to even come across the word neurodiversity. The idea is simple: everybody is wired differently. It’s why some of us like to talk a problem through while others want to chew on it alone, or why some people need an example before a concept clicks, and others find the example gets in the way. None of these is the right way - they’re just different ways. And what works for one person doesn’t necessarily work as well for another.
I’m now a heavy AI user, as are so many people, and so is everyone on our team. It became clear quickly to us – as a team focused on cognitive differences - that AI isn’t made for individual brains. It’s made for the average user. That’s fair; you have to start somewhere, and the AI tools of today are phenomenal in their efficacy. But in terms of how they fit with their human users, we hope they’re only at a starting point, and the next step is obvious.
(image of team at work)
Because the AI doesn’t know how you think, you do the work of getting on the same page. Shorter, please. Answer first. Give me an example. Not that tone. All that stuff. Then you do it all again tomorrow, especially as you add and try new tools, who don’t know or ‘get’ you off the bat.
We saw how common this is when we surveyed more than power AI users earlier this year. 76% told us they frequently rework what AI gives them, and about a third lose hours every week to it.
This isn’t about sycophancy. Nobody needs or should want an AI that agrees with them or flatters them. It’s about removing the unnecessary effort of getting on the same page.
You’ve probably had a colleague you just clicked with. Not because they thought like you, but because you gave each other room to work your own way. One of you wanted the headline first, the other wanted the backstory, and neither of you had to argue the point. That’s what you should get from your AI colleagues too.
That’s what Frequency is designed to do. It learns how you work with AI from how you use it, shows you what it learned, and lets you edit it, so answers fit you from the start.
If different is the norm, our tools should deliver there too.