AI tools are doing something specific to the developer-PM divide. Things that used to require years of specialisation are changing fast enough that the old categories — developer, product person, technical lead — don't sit cleanly anymore. I've been in software for over twenty years. I spent eight of them doing product. I'm not entirely sure which one I am now. This project is one way to find out.

What Proof of Support is

The app is a social tip jar. Visitors come to a page, leave a name and a message, and attach a small amount of the chain's native currency. The message goes permanently on-chain. The funds go to the owner's wallet — mine, as it happens. Which makes this something other than a demo. If someone finds the work useful enough to leave a message and attach something to it, the contract is ready.

It's simple enough to build in a week. It's complex enough to surface every meaningful difference between blockchain ecosystems: wallet integration, gas costs, frontend tooling, documentation quality, deployment friction. There's a reason I chose this over "hello world."

The format is a controlled experiment. The same app — identical UI, same contract interface, same three functions — gets built and deployed on a different blockchain each week for five weeks. After each deployment, I score the experience on an eight-dimension rubric covering developer tooling, contract authoring, documentation, frontend/wallet integration, deployment, transaction cost, and community support. The rubric was built before Week 1 starts, so the scoring is consistent, not retrofitted to whatever happened.

Five weeks, at least five chains. Categories: EVM L2, ZK-EVM, experimental EVM, non-EVM mainstream, and one wildcard slot — though the wildcard extends to the whole project. If a week surfaces something worth following, or feedback pulls in a direction, I'll add more. The structure is built for that. Chain selection is rolling — I'm picking each week's chain at the start of that week, based on the research, not in advance.

Who's doing this and why now

At some point I stopped being able to say "I built that." Not overnight — the transition is gradual. You're still close to the code, then you're reviewing it, then you're writing the brief for it. And one day you realise you've been in rooms where the technical decisions got made and you were the one asking what was feasible, not the one answering.

The landscape kept changing too. AI tools started doing things that used to require years of specialisation. The lines between developer and product manager started blurring in ways I hadn't expected. I found myself not quite sure which side of the table I should be sitting at — or whether that distinction still makes sense. This project is partly an attempt to find out.

I got burned. Real money in projects that promised yields, delivered quarterly newsletter updates, and then went quiet. I'd been in this space since the early mining days — be it BTC, LTC, FTC, whateverTC; then Android apps on Nxt when that was a serious network (I believe now with empty blocks it's not). None of that made me immune. It just meant I knew exactly what I'd walked into.

The return happened gradually: Solidity courses, a Lightning Network project, experiments on Gnosis... And then this — a structured experiment with AI as a collaborator, because that's the other part of the story. I'm explicitly using LLM tools throughout this project. Not to avoid the hard parts, but to see what the mode of "one senior plus AI pair programmer" actually produces.

What to expect, and what to do right now

Every week ends with a working live deployment — real app, real chain, accessible to anyone. And a retrospective: first-person, scored, specific. Not "great developer experience" but "here's the exact error I hit at 11pm and here's what it told me about the toolchain."

At least five weeks of this. I'll be posting on Farcaster (@satorigeeks) as it happens. The full retrospectives live here on Paragraph.

The series will also answer a different question for me personally: which part of this space is worth going deeper into. Which chain, which tooling, which kind of problem. The retrospectives are honest documentation — but they're also research for what comes next.

If you're a developer who's curious about blockchain development but suspicious of the noise, you're exactly who I'm writing for. Come back next week — the first chain goes live then, and so does the message wall. Be one of the first entries on it.

Week 1 starts soon. I've picked the chain. I'm not telling you yet.

AI tools are doing something specific to the developer-PM divide. Things that used to require years of specialisation are changing fast enough that the old categories — developer, product person, technical lead — don't sit cleanly anymore. I've been in software for over twenty years. I spent eight of them doing product. I'm not entirely sure which one I am now. This project is one way to find out.

What Proof of Support is

The app is a social tip jar. Visitors come to a page, leave a name and a message, and attach a small amount of the chain's native currency. The message goes permanently on-chain. The funds go to the owner's wallet — mine, as it happens. Which makes this something other than a demo. If someone finds the work useful enough to leave a message and attach something to it, the contract is ready.

It's simple enough to build in a week. It's complex enough to surface every meaningful difference between blockchain ecosystems: wallet integration, gas costs, frontend tooling, documentation quality, deployment friction. There's a reason I chose this over "hello world."

The format is a controlled experiment. The same app — identical UI, same contract interface, same three functions — gets built and deployed on a different blockchain each week for five weeks. After each deployment, I score the experience on an eight-dimension rubric covering developer tooling, contract authoring, documentation, frontend/wallet integration, deployment, transaction cost, and community support. The rubric was built before Week 1 starts, so the scoring is consistent, not retrofitted to whatever happened.

Five weeks, at least five chains. Categories: EVM L2, ZK-EVM, experimental EVM, non-EVM mainstream, and one wildcard slot — though the wildcard extends to the whole project. If a week surfaces something worth following, or feedback pulls in a direction, I'll add more. The structure is built for that. Chain selection is rolling — I'm picking each week's chain at the start of that week, based on the research, not in advance.

Who's doing this and why now

At some point I stopped being able to say "I built that." Not overnight — the transition is gradual. You're still close to the code, then you're reviewing it, then you're writing the brief for it. And one day you realise you've been in rooms where the technical decisions got made and you were the one asking what was feasible, not the one answering.

The landscape kept changing too. AI tools started doing things that used to require years of specialisation. The lines between developer and product manager started blurring in ways I hadn't expected. I found myself not quite sure which side of the table I should be sitting at — or whether that distinction still makes sense. This project is partly an attempt to find out.

I got burned. Real money in projects that promised yields, delivered quarterly newsletter updates, and then went quiet. I'd been in this space since the early mining days — be it BTC, LTC, FTC, whateverTC; then Android apps on Nxt when that was a serious network (I believe now with empty blocks it's not). None of that made me immune. It just meant I knew exactly what I'd walked into.

The return happened gradually: Solidity courses, a Lightning Network project, experiments on Gnosis... And then this — a structured experiment with AI as a collaborator, because that's the other part of the story. I'm explicitly using LLM tools throughout this project. Not to avoid the hard parts, but to see what the mode of "one senior plus AI pair programmer" actually produces.

What to expect, and what to do right now

Every week ends with a working live deployment — real app, real chain, accessible to anyone. And a retrospective: first-person, scored, specific. Not "great developer experience" but "here's the exact error I hit at 11pm and here's what it told me about the toolchain."

At least five weeks of this. I'll be posting on Farcaster (@satorigeeks) as it happens. The full retrospectives live here on Paragraph.

The series will also answer a different question for me personally: which part of this space is worth going deeper into. Which chain, which tooling, which kind of problem. The retrospectives are honest documentation — but they're also research for what comes next.

If you're a developer who's curious about blockchain development but suspicious of the noise, you're exactly who I'm writing for. Come back next week — the first chain goes live then, and so does the message wall. Be one of the first entries on it.

Week 1 starts soon. I've picked the chain. I'm not telling you yet.