Why this exists

I have been shipping software on my own since 2015. This is the first thing I have built where the hardest engineering decision and the hardest moral decision turned out to be the same decision.

The idea started somewhere unremarkable. I wanted to know what was in my own backyard at 2:00 AM without opening the door. Cameras give you four angles and a ten second clip of a shape. That is not knowing. That is being told something happened.

Building a drone that flies a loop and streams to your phone is a solved problem. I could describe the parts list in a paragraph. What stopped me was reading how the last company to build serious autonomous security tech ended up being used.

Flock Safety built a camera network, wrote a policy prohibiting immigration enforcement use, and then roughly 4,000 immigration searches ran through it anyway. Not because they were villains. Because they built a capability, wrote a rule about it, and the rule turned out to be the weaker object. 214 cities and counties have since ended their contracts.

Immigration searches: 404 Media, 2025. Contract cancellations: DeFlock tally, reported August 2026.

I read that and understood the actual product problem. Anyone can build the drone. The question is what stops the drone from becoming the thing people are ripping off poles in four years.

The answer is that you have to give something up, publicly, before anyone asks you to.

So the model runs on the dock and not on my servers. I hold no keys, which means I could not hand over your footage if a court ordered me to, if an acquirer demanded it, or if I simply changed my mind in five years and decided the money was worth it. That last one matters most. I am designing against my own future self, because he is the threat model nobody writes down.

The neighbourhood mesh is the obvious next product and I am not building it. Every home with a drone, sharing what it sees, a network across a subdivision. It is worth more than everything else in the plan combined. It is also precisely the thing this company exists to be the alternative to, and a promise I have not tested with real money in front of it is not worth much. So I am declining it now, in writing, dated, while it is still theoretical and cheap to decline. The nine commitments are at /commitments/v1, versioned, so you can hold me to the exact wording later.

I should be straight about what I am not. I am not an aerospace engineer. I have shipped software for over a decade and I run several consumer brands, and the aircraft will be built by an American manufacturer who does this properly. What I am bringing is the software, the dock, and a set of refusals I intend to be held to.

If you have spent this year arguing with your city council about cameras, or you just want to see your own backyard without getting out of bed, we want the same thing. Take a position in line.