The rise of the programmer-producer

During the Summer Cohort in July of this year, we got into a discussion on the future of programming and what’s coming next. I can’t remember who said it, but I’ll never forget it:

If we can build anything now, why don’t we? But for ourselves…

That’s the summary of the thought, and I’ve been pondering it ever since. If we can build it ourselves, why do we need clients to dream it up first? This is the entire theme of the Fall Cohort and I’ll discuss more on that in a minute.

Anyway: I’ve started calling the person who does both the programmer-producer: you write the code and you also decide what gets built. I think a lot of us already have that job, and building Kolea is how I found out I do.

A long-running dream

I know I’ve mentioned more than once that I’ve had a dream for a very long time of running my own email list service. It’s a complex problem, and the idea of tackling it and “getting it right” has been my nerd-need for quite a few years. A few months ago, I decided to take Fable for a spin and I ended up building the exact thing I’ve wanted for a very long time.

I open-sourced it, of course, and since then have sent 10s of thousands of emails through it with only some slight bugaboos that had more to do with Cloudflare than anything else (which ultimately ended up being my fault). I kept adding features too, including a templating system that made it feel like Ghost, and a fulfillment system that made it feel a bit like SendOwl, the digital fulfillment service I paid a lot of money for once upon a time.

That’s when I began to think about creating a hosted service, a “pro” version if you will. I’ve wanted this since 2012, before I sold Tekpub, when I was paying over $300/month to stitch together email, fulfillment and a store. Just a few months ago I was paying almost $600/month for these services and, don’t get me wrong, you have to pay to play in this space.

But then again, with AI, I can make my own damned space can’t I?

Who needs clients when you can build the thing yourself?

So here it is: I made a pro version of Kolea, which is the very first SaaS app I’ve made. I did 100% on my own with Claude Fable and Opus, and it’s easily the greatest piece of software I can hang my name on.

I was having so much fun throwing this thing together that I just couldn’t stop myself. Partway through, I scrapped the React/Bun stuff and went with Elixir so I could lean on the BEAM resilience and speed. With a client, that rewrite would have been a meeting, an estimate and probably a no. With Claude it took me three nights, with each night clearing off a queue of issues at GitHub where each one was a prompt of a feature that needed building.

This is the reverse of the Spolsky rule: when a rewrite is that cheap, the question stops being whether you can build it, how much, and how long. It becomes a question of “how many tokens do I have left in the tank” and which model to use.

And I know that’s the question you want answered right now: how much did this cost to build? The answer is nothing on top of what I was already paying Anthropic. I have my $200/mo Max Plan, and that’s precisely what I used. I didn’t have loads of agents going off at once, I had one orchestrator running builds for me at night and during the day as I recorded the next drop of The AI Pro.

Which brings me back to the question from the cohort. If you can build it yourself, what do you actually build? A replacement? Or something better?

The answer is obvious! Every app you make from now on needs to be a piece of wonderful art that not only functions and scales better than anything you’ve made to date, it also has to look the part too.

Above all: make it gorgeous

This is the thing that probably took most of my time: fine tuning the interface to look amazing and make good sense. Writing is hard! Knowing what to write about is even harder still, and I think the current email services out there absolutely suck at helping move past the dread of the blank page.

I daresay I got around that with Kolea. Yes, it’s AI helping you think, but it’s “Just in Time” AI that doesn’t do too much. You can use your own keys (Anthropic, OpenAI, or OpenRouter) or use mine if you sign up for a pro account.

I love the way it looks. I love the way the dashboard makes me feel like a Saturday afternoon at the beach (which is part of the prompt I gave it). I absolutely love this pallete!

When you visit your dashboard, I want you to know immediately what’s going on with your newsletter and comms site. Who’s coming in, from where, and what they’re reading. I want you to see their journey from your site through to a possible sale (if that’s your gig) or some other conversion.

I could go off on the numbers we’re tracking, but I’m going to stop and also add that no, there is no tracking outside the app at all. We read incoming requests, track clicks and opens in the emails we send, and whenever a sale happens at Stripe (via webhook). That’s it!

What I didn't expect: I have a new job

Five years ago, an app that takes the place of four established SaaS products would have meant a team, a bunch of funding, and probably a few years of my life full of stress and anxiety. This time it was me, a microphone (I use Wisprflow), and a fleet of agents that write code faster than I ever could and do a really good job of it.

I had nothing but fun, and would shout, at least daily: “you have to come see this OH MY GOD” across the house.

A new kind of job is taking shape, right before our eyes, and I think a lot of us already have it without knowing it. I've been calling it the programmer-producer, and at the core of it is a simple question:

Is it easier in the long run for a business ideas person to learn to build applications with AI, or for us to learn how to create amazing businesses?

I’ve done both, if I’m honest, and the difference between them is the biggest thing I learned building Kolea. I’ve been on the client end of things with Tekpub, when you hire a site builder and designer only to scratch your head in amazement at the complete lack of care and inspiration.

I’ve also been in the builder’s seat, blown away that a client would rather see a spreadsheet interface on a web page rather than a gorgeous informational card! Just think of the readers, dangit!

I think it’s absolutely clear which case is better here: ours. We know what people want with these apps because we’ve built them so many times and had to support them. We know the complaints, the pain points, and now we can address all of them and more, but faster.

I think a lot of the SaaS systems built up over the last 10 - 15 years are about to age out. One person with agents can now write, test, refactor and fix code at a pace that used to take a team weeks or months, and the result will be superior because its one person’s vision, not 5. This will reshape how software gets made in the next 5 or so years, and I’m excited about it.

Kolea is my statement on this. Kit, Ghost, and even Shopify can’t touch what I’ve just made, and they absolutely can’t fix, tweak, improve, and add features like I can. Kit alone will cost you roughly twice as much for an email list service that does 10% or less of what Kolea does. Fulfillment services will cost you even more just to track downloads and create expiring URLs.

Ghost Pro will give you a dandy CMS system with basic analytics, no review process and no notion of whether your writing is moving a needle. They don’t have automated sequences, and your landing pages conform to their templating with no MCP server to speak of.

This isn’t meant as a hit on these services - they’re just built on and tied to a business model from the past, and pivoting right now means disruption and overhaul. I get to start from scratch, flexing my experience as both a client and a builder, creating something beautiful. Even if I’m the only person to ever use this thing, it’s completely worth the effort to me.

And that’s where we are: a generation of talented, experienced programmers armed with agents that are going to remake the online experience by themselves, or with very, very small staffs. So what would you build if nobody had to hand you a spec first? I’d really like to know.

If you want to try it

First, Kolea is ready for invites. If you, like me, are about to build your Big Thing. It can help you get the word out and, more importantly, help you know if what you’re doing is working. If you don’t know what you’re doing - that’s even better. All the help you need is right inside.

The second thing is: if you want to kick off your side gig or Next Big Thing with a group of folks, that’s what the Fall Cohort is all about: LEVERAGE. Flexing your know-how as a senior programmer to build something incredible, all on your own.

If you want to know more, you can read up in it here. I have exactly 5 seats left, so jump on in. This is going to be fun!

As always: thanks for reading and yes, I wrote this myself despite having a Ghostwriter next to me :).

Rob

Sent with love via Kolea.