You've sent a photo to ChatGPT.
Maybe it was a screenshot of a document. A chart you needed explained. A picture of your whiteboard after a long meeting.
Or this one?
And it answered. Accurately. Maybe even impressively.
So here's the question nobody asks:
What did the AI actually see?
The answer is going to bother you a little.
Nothing.
It saw nothing.
Not in the way you see things. Not even close.
Here's what actually happened.
The moment you uploaded that image, it got converted into numbers. Thousands of them. Each number represents a pixel — its colour, its brightness, its position on a grid.
Your photo of a whiteboard? To the AI, it arrived as something closer to a very long spreadsheet! 😅
The AI then ran those numbers through a series of mathematical calculations — billions of tiny weighted decisions — and produced an output. Text. An answer. A description.
It never "looked" at anything.
It processed. It predicted. It responded.
This is the part most people miss about AI entirely.
AI is not intelligent the way you are intelligent. It doesn't understand. It doesn't "reason". It finds patterns in numbers and predicts what the most statistically likely next output should be.
Your image goes in as numbers. Your answer comes out as numbers. The interface just makes it look like magic. ✨
Why does this matter?
Because once you understand what AI is actually doing, you stop being impressed by the trick and start seeing the tool.
And tools you understand, you can use better than everyone around you who's still just watching the magic show.
That's exactly what separates PMs who use AI to save two hours a week from PMs who use it to restructure how they work entirely.
One group sees a magic trick. The other group knows where the rabbit is hidden.
If you want to go from the first group to the second, I put together something that will get you there fast.
It's called the AI Accelerator — built specifically for project managers who want to understand AI well enough to use it at a level most people in their org never will.
No fluff. No theory for theory's sake.
Check it out here →
Got questions or comments? Let me know. I read everyone!
John
Responses