What if the problem isn't the AI — but you?
Imagine you're trying to build a LEGO castle, and you keep running into problems. You ask your friend for help, and they say, 'Let me just build it for you.' But then you realize, the real issue isn't that your friend can't build it — it's that you didn't think about all the pieces you needed. That’s the core idea behind a new approach to working with AI, especially with powerful models like Claude 5 (also known as Fable 5).
What is a blind spot in AI work?
A blind spot is something you don't know you don't know. It's like not realizing you're missing a key ingredient in a recipe, or not understanding how a particular part of a machine works. In the context of AI, a blind spot refers to the knowledge or thinking gaps that you have — but don't even realize you have.
When working with AI tools, people often assume the AI is the problem — that it’s not smart enough or not doing what they want. But sometimes, the real issue is that we haven't thought through our own thinking first.
How does this work with AI?
Think of it like this: if you're a chef, and you're trying to make a new dish, you might start by asking an AI for a recipe. But before you do that, it helps to ask yourself: What ingredients do I actually have? What do I not know about this dish? What might I be forgetting?
This is exactly what a blindspot pass is. It's a method where you go through your own thinking step by step — asking questions like:
- What do I think I know?
- What am I not sure about?
- What assumptions am I making?
- What might I have missed?
By doing this, you help yourself realize what you need to tell the AI — or even what you might need to learn before you even ask.
Why does this matter?
When you work with AI tools, especially ones as advanced as Claude 5, the AI can do a lot. But it still depends on what you tell it. If you don't think through your own blind spots, you might give the AI unclear instructions or miss important details. This can lead to results that are not quite what you wanted.
By using techniques like blindspot passes and structured interviews (which are like asking yourself a series of thoughtful questions), you can make your AI interactions more effective. It’s not just about asking the AI to do the work — it’s about making sure you're ready to ask the right questions.
Key takeaways
- A blind spot is something you don't know you don't know — and it can affect how well AI tools work for you.
- Before giving a task to an AI, it helps to think through what you might be missing or assuming.
- Techniques like blindspot passes help you uncover these hidden gaps in your knowledge.
- Working with AI is not just about the AI — it's also about how well you prepare and think through your own thinking.
So next time you're working with an AI, try pausing first. Ask yourself: What do I know? What do I not know? And what might I be missing? You might be surprised at what you find.
