A few moments ago, I came across an intriguing post on Reddit that posed a simple yet provocative prompt: “Now that you can remember everything I’ve ever typed here, point out my top five blind spots?” (source). The idea stuck with me.
As someone who’s used ChatGPT extensively across a wide range of projects—academic writing, software development, teaching, and even a bit of whimsy—I figured the model might actually have something insightful to say. After all, it’s been part of my daily workflow for months, across everything from syllabus edits to tour app debugging to conversations about cats and Canadian geese.
So I asked.
Here’s what it said.
My Top 5 Blind Spots (According to ChatGPT)
1. Tendency to Overbuild Before Testing
I often refine systems and features to a high level of polish before seeing how they work in the wild. This can slow down feedback and iteration.
2. Taking on Too Much Solo
I frequently carry the full load of a project myself, even when collaboration could make things easier—or better.
3. Over-Correcting for Clarity
My preference for precision and structure can sometimes get in the way of creativity or navigating ambiguity.
4. Fast-Paced Thinking, Slower-Paced People
I move quickly between ideas, modes, and tasks. Others don’t always keep up, and I don’t always pause to bring them along.
5. Understating My Influence
I tend to frame my statements cautiously, even when I have the authority or expertise to speak more definitively.
I have to say, this was more insightful than I expected. These aren’t things I’m completely unaware of—but they’re definitely themes I don’t always recognize in the moment. It’s a little like holding up a mirror that reflects not just how I act, but how I tend to act across contexts.
And that, honestly, feels like one of the most useful ways to use tools like ChatGPT—not just as a productivity enhancer, but as a thinking partner that notices patterns you might miss. In some cases, I can clearly see where some of these blind spots come from – for example, when you have a lot of interests, it can be hard to find collaborators because you don’t get deep into communities around one interest. And it doesn’t help when they may move too slow for you anyway (see #4)!
I’m not sure what, if anything, I should do about this. But it is interesting pushing an LLM to be critical of you and provide constructive comments. It’s a far cry from the usual flattery it provides. And perhaps useful if for nothing else than to see what it’s been thinking about you behind your back!