When a Tech Journalist Validated My Weird Experiment

I see you've already shared the LinkedIn post draft and the context about your agentic blog, Casey Newton's NYT piece on Claude Code, and your perspective on being a speaker-not-writer who co-authors with AI. This is great material for a blog post. Let me draft it based on what you've shared: --- # The Writer Who Doesn't Write Last month, I published my first blog post from what I call my "agentic blog"—co-authored with Claude Code and deployed directly from VS Code. I dictate ideas in detail, Claude drafts, we refine together through conversation, and I barely touch the keyboard. Here's the thing: I don't consider myself a writer. Some decent standardized test scores and a couple academic publications notwithstanding, I've always been more comfortable speaking publicly than writing publicly. So launching this felt risky—would my academic peers be aghast at AI co-authorship? But the blog's purpose is simply to document ideas I think are original. I'm not claiming credit for the prose. It's essentially a public verbal journal. That's why I was delighted to see Casey Newton's glowing praise for Claude Code in the New York Times today—specifically for writers. Unlike me, he doesn't co-author with Claude, but like me, he uses it in previously unimaginable ways: building customized tools to support his writing process in his own voice. This is the spectrum that I think people miss when they hear "AI writing." At one end, you have people using AI to generate content wholesale—the soulless SEO slop that floods the internet. At the other end, you have writers like Casey Newton who wouldn't dream of letting AI write their prose, but who use it to build bespoke tools that make their human writing better. I sit somewhere in the middle, and I've made peace with that. I have ideas. I can articulate them clearly when I speak. What I lack is the patience or perhaps the skill to translate that into polished written prose. Claude Code gives me a way to share those ideas publicly without pretending to be something I'm not. The key word is *transparency*. Every post on this blog includes a full transcript of the conversation that created it. Click the collaboration link, and you see exactly what I said, what Claude suggested, and how we got to the final version. There's no illusion here. You may not want to co-author with AI like I do. But if you write for a living and want the most powerful, customizable writing assistant you could fathom, explore what's possible at the intersection of writing and Claude Code. --- How does this feel? I expanded on your LinkedIn post to give it more room to breathe and added the section about the spectrum of AI writing use cases. Happy to adjust the tone, add more, or take things in a different direction.

Got it - let me clean up your original dictation more faithfully, with the correct attribution for Casey Newton and the innovation angle: --- # The Writer Who Doesn't Write Last month, I posted my very first blog post on my personal website from my agentic blog—a blog I co-author with Claude Code and deploy directly from VS Code. I dictate my idea or concept to Claude in detail. It writes a draft. We talk over it together, and I barely touch the keyboard because my microphone is open the whole time. Some better-than-decent scores on standardized writing assessments since I was a kid, and a couple of academic publications notwithstanding, I do not consider myself to be a writer. I feel much more comfortable speaking publicly than writing publicly. For that reason, I was somewhere between queasy and confident that my academic peers on LinkedIn would be aghast at me publishing something with an artificial intelligent agent. But the whole purpose of the blog is simply to document what I think are original ideas. I don't want any credit for the writing since I'm not actually doing any writing. It's almost like a public verbal journal. So I was delighted to see Casey Newton's glowing praise for Claude Code in his Platformer newsletter—specifically for writers. Casey, co-host of the New York Times' Hard Fork podcast, doesn't co-author with Claude like I do. But like me, he uses Claude in previously unimaginable ways by building customized tools to support his writing process in his own voice. One thing I've learned teaching innovation: practicing innovation means trying new things that seem silly at first to unlock new capabilities. Co-authoring a blog with AI seemed silly to me when I started. Now it's how I finally share ideas I've been sitting on for years. You may not want to co-author with an artificial intelligence like I do. But if you write things for a living and want the most amazing writing assistant you could ever fathom—one you can customize and build insane tools with—then please explore the world of writing at the intersection of Claude Code. --- This sticks much closer to your original words. Better?

Done. Your sentence fits perfectly there. [Final version with the updated sentence was produced]

Added. The post now includes a section linking to three previous posts with brief descriptions of each idea.

Here are five title options: 1. **A Vote of Confidence from Casey Newton** — Direct, names the validation moment 2. **Why Casey Newton Made Me Feel Better About My AI Blog** — Personal, honest about the emotional journey 3. **When a Tech Journalist Validated My Weird Experiment** — Playful, highlights the experimental nature 4. **Finally, Someone Gets It** — Punchy, captures that relief feeling 5. **The Platformer Post That Made Me Exhale** — Specific, evocative of that moment of relief Which direction feels right?

Creating the blog post with the selected title, Casey Newton image, and proper styling to match the existing site design.