The artifact mindset: building in public with Claude

It started as a small experiment, almost a dare to myself: what if I shared every interesting thing I made with Claude — not the polished portfolio pieces, but the raw artifacts? The half-finished tools. The prompt that finally clicked after twenty that didn't. The ugly first version I was a little embarrassed by.

I expected silence. What I got instead were conversations. People didn't just like the posts on LinkedIn — they replied, asked how I'd gotten Claude to do a thing, and started posting their own experiments back at me. Somewhere in there my whole idea of what's worth sharing quietly flipped.

What I actually mean by an “artifact”

An artifact isn't a product. It's the thing that exists between an idea and a product — and most of us are trained to hide that middle. For me an artifact is any of these:

• A Claude conversation that solved a problem in a way I didn't expect.
• A tiny tool I built in an afternoon that does exactly one thing well.
• A prompt I iterated on until it became reusable — and the dead ends that got me there.
• A generated draft, mockup, or script that's 80% done and clearly imperfect.

None of those feel “done.” That used to be my reason not to post them. Now it's the reason I do. The unfinished state is the interesting part — it's where the thinking is still visible.

Polished work hides the part people want

When I ship something clean, people see the result and nod. When I ship the artifact — the messy middle — they see how, and that's what they actually came for. Building in public with AI tools is a different kind of vulnerability than open source. You're not just showing your code. You're showing your thinking, your prompts, the order you tried things in. It's more intimate, and that intimacy is exactly what makes it resonate.

Polish signals “look what I made.” An artifact says “here's how I'm thinking — come think with me.” Only one of those starts a conversation.

There's a fear to get past first. Posting a prompt feels like handing over the recipe; posting a half-built tool feels like inviting people to point at what's broken. What I learned is that almost nobody shows up to point. They show up to add. Someone always knows the next move, the better phrasing, the edge case I missed — and they tell me, for free, because I left the door open.

The doors it opened

“Opened doors” sounds grand, so let me be specific about the unglamorous, real version. Sharing artifacts got me into conversations with people far more experienced than me. It turned passive followers into collaborators who'd DM me “you should try X.” A few of the rough things I posted turned into the experiments and tools on this very site — they stopped being throwaway artifacts and became real, because sharing them gave me a reason to finish.

None of it came from a viral post or a clever growth tactic. It came from showing up consistently with honest work-in-progress. Compounding, not luck.

How it changed the way I build

The biggest shift wasn't external — it was in how I work. When you know you'll share the artifact, you build in smaller, sharper units. You stop hoarding a giant unfinished project and start shipping the one useful piece. The build → share → learn loop got tight enough that the sharing started feeding the building.

It also made me kinder about my own first drafts. You can't post the messy middle for months and keep believing your work has to be perfect to count. The artifact mindset is, quietly, a cure for the kind of perfectionism that keeps good ideas in a drafts folder forever.

If you want to try it

A few things I'd tell anyone starting:

• Post the artifact, not the write-up. Share the actual prompt, the actual tool, the actual conversation — not a polished essay about it later.
• Show the prompt. It's the most generous and most interesting thing you can include, and it's the part people quietly screenshot.
• Ship before you're ready. “Here's a rough version” outperforms “one day I'll post the finished one,” which is to say it outperforms never.
• Narrate the iteration. The twenty attempts that failed are more useful to people than the one that worked.

I'm still doing it. Most of what you see across this site started as an artifact I almost didn't share. The mindset is simple, even if the first post never is: make the thing, then let people watch you make it. The work gets better, and you're not building alone anymore.