Every essay here started out as a conversation. Not a prompt I typed in and walked away from, but an actual back-and-forth, the kind that wanders around a topic for a while before it figures out what it's actually about. My usual partner in those conversations is Claude, an AI assistant made by Anthropic. So the name of this site isn't really a joke. It's just what happened.
I'm Iko Knyphausen. I live in Lake Forest Park, Washington, but I grew up in Europe and spent thirty years there before landing in Seattle, which means I watch American politics with one foot still somewhere else: close enough to care, far enough away to notice the things everyone here takes for granted. I used to run a software company; now I'm retired and spend my time on photography, travel, and writing about the stuff that's been bugging me. I write about policy because that's usually where the gap between what a country says about itself and what it actually does shows up most clearly. And that gap is almost always bigger than people expect.
Here's roughly how it goes. Something's been bothering me, or I've been reading about it, or arguing with myself about it in the shower. I bring it to Claude. We dig through it together: the evidence, the angles, the best counterarguments. Claude searches, summarizes, occasionally tells me I'm wrong. I push back, redirect, insist on something. Eventually an essay starts to take shape, and I write it, and Claude edits it, and I revise it again, and what comes out the other end is something neither of us would have come up with alone.
I'll be upfront about what that means in practice. The research is real, the sources check out, and the arguments are ones I actually believe and will defend if you push back. The sentences and the voice are mine. What Claude brings to it is harder to put into words: a kind of patient, thorough engagement that doesn't get tired, doesn't have an ego riding on the outcome, and is genuinely useful to argue with.
People sometimes ask if using AI to write makes the whole thing somehow fake. I've thought about that more than once. What you're reading is still the product of my own reading, my own judgment, my own opinions about what matters and how to say it. Claude's a genuinely useful tool, but it doesn't actually care whether drug policy gets reformed or how a court should have ruled. I do. That caring is the essay. The AI just helps me get it onto the page in better shape than I'd manage solo.
If that bothers you, I get it. These are real, unresolved questions about what authorship even means right now. I don't think I have a clean answer. But pretending AI isn't part of how a lot of serious writing gets done these days feels like a bigger lie than just telling you how the sausage gets made.