Electric Dreams of (automatic) Enterprise Software

I used to absolutely love coding when I was a kid. It was just this massive creative space where I could do whatever I wanted (with enough nested if statements thrown in). To me it was the most fantastic idea: imagine something, make something, have fun, games, NPC character generators for Dungeons and Dragons, databases to manage supply chain stuff for scouts (yes ok I was a weird kid).

As I got older and the companies I had founded grew from a scrappy crew of pirates to big global enterprise software companies, I realized how real code had to be: debugging, object oriented, tests, infrastructure, git, and all that stuff. It wasn’t that I couldn’t get it, it was the fact I didn’t want to do it. It wasn’t fun. That was when I realized I suck as a software developer. The distance between idea, iteration, payoff, and then scaling was simply too far. Even though the companies I co-founded actually built massive enterprise systems (thanks to way smarter people than me), my personal coding sweet spot was always solving my own creative needs. I still kept pushing myself to learn, try new technologies, fail, and try again, but the days of me having AWS certificates on my laptop for production are long gone. I’m still a nerd at heart, but getting from idea to production feels like a big mountain, and I know my time is better used elsewhere.

Then slowly, first and now faster and faster, something happened. Every 6 months I’ve been using whatever the state of the art is for codegen to make stuff and to see how far I can get before the layer cake collapses. It’s great fun and, I think, a good way to really understand how AI works and what its limitations are. But recently the layer cake stopped imploding and instead kept getting better and actually worked. I was able to go from idea to execution of even very complex systems and manage them. Instead of making infrastructure for debugging, deployment, and all that boring stuff, I found I had created my own infrastructure for creativity. Using GitHub repositories for templates, Cursor and Cline for development, MCP servers for testing and code review, Docker for containers. The best part? I rarely touched the code. With Composer in Cursor, Claude Code, and my own built tools and agents, I could finally get stuff done.

I work with a team of some of the most brilliant developers I’ve ever worked with in my life. In my excitement I started talking to them about my new workflows: how I was working, using deep research to pull context for a new experiment, using AI templates for managing projects in Cursor, tracking progress, creating rules for the composer agent to code the way I wanted, all with the purpose of taking the engineering out of software. They just didn’t see it that way. Being engineers and knowing how to code, they could never dream of going to the lengths I did, not to code.

This got me thinking. I think we are looking at what AI means for software development all wrong, because for every hardcore software developer with the mind and stamina to go through the valley of real engineering (and I salute you) there are 100 more like me. Software development is power, it’s the power to turn your ideas into reality, and in a world where most things and services and value are virtual, that’s a superpower. I think we’re just at the beginning of the tsunami of change (good and bad) this will unleash.

Now the cool thing about Beyond Work is of course that this is exactly what we’re building: the ability for anyone (in business, for the other weird kids like me dreaming of solving enterprise problems) to build whatever they want to solve the problems they have, in a safe, compliant and secure manner of course. The last time something like this happened was when Excel got macros.

Enterprises need their lovable software too.

Ps. A few people asked me for my current workflow, this is what I do currently: