A declarative home base

I created my personal website mdwt.dev today - something between a portfolio and a declarative home base. It's a response to how my reality has changed over the last few months.

The reality of flying solo is that, after building in isolation for so long, I tend to forget some important things I've done and the lessons and mistakes that have shaped my journey as a developer and an entrepreneur. My LinkedIn page sums up the last few years as "had a job until 2019 and then ran a CheckoutJoy until 'present'" - which, like Bret St. Hubbins describing the 18" Stonehenge, tends to understate the hugeness of the object.

Right now, everything I do is a secret shared only between me and my coding agents (the --dangerously-skip-permissions kind). But that magic we share only lasts until I switch to a different model, or they start clanking and I clear context, and then I'm on my own again. Users see the result of the work, sure enough, but the process of getting there is hidden and mostly lost. I've been churning out code for a long time - features, funnels, schedules, bots, platforms, UIs - real code out there in the wild with real outcomes.

So now feels like the right time to acknowledge the work I've done, the choices I've made, and what I've achieved along the way. It's also a chance to explore where my expertise and skills fit in.

checkoutjoy.com is still my main focus, and something I need to see through. It's been a part of my life for 6+ years so far, days and nights, weekends and holidays. That's long enough to reflect a real part of who I am (it embodies everything I love and everything I hate doing), but it doesn't define me now like it used to. For now, I'm celebrating small victories and relishing the humble pie along the way.

I'm also building getpaidhq.co - a subscription payment and dunning engine in an ongoing attempt to solve proper subscription billing for people outside of the Stripe multiverse - it's still one of the main pain points I see every day working with CheckoutJoy sellers.

Then there's Keepem.io - a usage-billed webinar platform. Something about video, combined with AI and the processing/measuring/billing use case, intrigued me so much that I spent real time and plenty of tokens designing and building it. Positioning is still a mystery as I'm not quite sure what problem it solves yet, but it's an interesting and growing space and always in the back of my mind.

I've also been on this mission to replace some systems I use in my business - the kind of things that seem easy to do but never really justify burning focused development hours on. Luckily, in 2026, that whole genre of thinking is a thing of the past. mailshot.dev is one result of this - an AI-first open-source framework for sending marketing and onboarding emails with AWS. I built it to stop paying SaaS rent for something silly like email contacts and sequences. It's as enjoyable to build as it is to use.

Everyone is still figuring out where AI fits in, if and when AGI will become a thing, and what parts of our lives we're willing to hand over to AI. I lean on a few YouTube heroes like Mo Bitar to help me make sense of whatever this is we're going through. So far, all we know is that between the total enshittification of everything we've ever valued and AI being truly incredible as we're promised, the truth sits somewhere in the middle.