I came to Claude the same way most people do — with a task. I had content to write, ideas to organize, and a very specific belief that I would type something in and get something useful back. I would review it, edit it, maybe tweak the tone. Like a faster, cheaper assistant.
That's not what happened.
The first real moment of friction came when I asked Claude to help me write a post about my SketchUp course. I gave it everything—the audience, the topic, the vibe—expecting it to mirror my creative process. What came back was technically correct and completely wrong. It had my information but not my thinking. It sounded like someone who had read about me, not like me.
My first instinct was to fix the output. I started editing sentences. Ten minutes in, I realized I was rewriting the whole thing. The tool hadn't failed — I had used it wrong.
So I tried again, differently. Instead of giving Claude a task, I started explaining my thinking. What I actually believe about design education. Why I built the course the way I did. Who the student is that I keep imagining when I make decisions. I wasn't asking for output anymore — I was thinking out loud to something that would respond.
That's when it shifted.
Claude didn't just reflect my words back at me. It asked the question I hadn't asked myself. It pointed to the gap between what I said my course was about and what I was actually describing. That's not autocomplete. That's not a search engine dressed up in a chat window. That's something closer to a thinking partner — and that framing changed everything about how I work with it.
I want to be specific about what I mean, because I’m not making a grand claim about AI for designers. Claude doesn't have taste. It doesn't have experience. It has never sat in a client meeting, never struggled with a brief that kept changing, never looked at a design and felt — in the body, not just in the head — that something was off. Those things are mine. I bring them.
What Claude does is help me be more precise about things I already know but haven't fully articulated. It catches contradictions. It can hold a long, complex conversation about strategy without forgetting what I said ten minutes ago. When I'm stuck, it doesn't panic with me — it just keeps working. For someone who overthinks and moves slowly because of it, that consistency is genuinely useful.
I still edit everything. I still bring the judgment. But I stopped thinking about AI as a tool that does things for me, and started thinking about it as a place where I go to think.
That's not what happened.
The first real moment of friction came when I asked Claude to help me write a post about my SketchUp course. I gave it everything—the audience, the topic, the vibe—expecting it to mirror my creative process. What came back was technically correct and completely wrong. It had my information but not my thinking. It sounded like someone who had read about me, not like me.
My first instinct was to fix the output. I started editing sentences. Ten minutes in, I realized I was rewriting the whole thing. The tool hadn't failed — I had used it wrong.
So I tried again, differently. Instead of giving Claude a task, I started explaining my thinking. What I actually believe about design education. Why I built the course the way I did. Who the student is that I keep imagining when I make decisions. I wasn't asking for output anymore — I was thinking out loud to something that would respond.
That's when it shifted.
Claude didn't just reflect my words back at me. It asked the question I hadn't asked myself. It pointed to the gap between what I said my course was about and what I was actually describing. That's not autocomplete. That's not a search engine dressed up in a chat window. That's something closer to a thinking partner — and that framing changed everything about how I work with it.
I want to be specific about what I mean, because I’m not making a grand claim about AI for designers. Claude doesn't have taste. It doesn't have experience. It has never sat in a client meeting, never struggled with a brief that kept changing, never looked at a design and felt — in the body, not just in the head — that something was off. Those things are mine. I bring them.
What Claude does is help me be more precise about things I already know but haven't fully articulated. It catches contradictions. It can hold a long, complex conversation about strategy without forgetting what I said ten minutes ago. When I'm stuck, it doesn't panic with me — it just keeps working. For someone who overthinks and moves slowly because of it, that consistency is genuinely useful.
I still edit everything. I still bring the judgment. But I stopped thinking about AI as a tool that does things for me, and started thinking about it as a place where I go to think.