We design and ship a small family of focused products, led by GS Keystone and alongside Magpie. The tools change; the way we approach building them doesn't. A few principles guide every product decision we make.
Problem first, technology second
We start with the workflow we're trying to fix, not the stack we want to use. It's tempting to reach for whatever's interesting at the moment, but a product built around a trend outlives the trend rarely. We'd rather understand exactly where a team loses time or data, then pick the simplest technology that solves it well.
That discipline shows up in small decisions constantly: whether a feature needs a new database table or can live in an existing one, whether a workflow needs a new screen or can be folded into one that already exists. Every addition has a maintenance cost for as long as the product is alive, so we ask whether the problem is real before we ask how to solve it.
Ship software that lasts
Boring, maintainable systems beat clever ones. A clever solution is fun to build and expensive to keep running — someone has to understand it six months later, usually a different someone than the one who wrote it. We build to be useful years from now, not just for the next release.
In practice, this means favoring plain, well-understood approaches over novel ones unless the novel approach earns its complexity. It means writing code that's readable by someone who didn't write it. And it means treating a shipped feature as the start of a maintenance commitment, not the end of a task.
Honest about tradeoffs
Every product decision has a cost — speed against thoroughness, flexibility against simplicity, doing it now against doing it right. We surface that cost and name it, rather than let a customer discover it later. If a feature will take longer because we're building it to hold up, we say so. If a shortcut means revisiting something down the road, we say that too.
We think this is the only version of "move fast" that's honest. Moving fast at the expense of a customer's data, workflow, or trust just defers the slowness, and adds interest.
What this adds up to
GS Keystone runs real commercial solar projects today. Magpie holds real notes and ideas for the people who use it daily. Neither is a demo or a prototype dressed up for a pitch. They're working tools, built the same way we'd want a tool we depend on to be built: understood before it's started, made to hold up, and honest about what it costs to get there.
If that's the kind of software you're looking to build with — or the kind you're looking for as a customer — we'd like to hear from you.