Don Moskaluk

Modern boats know their engine temperature, fuel burn, battery state, location, heading, wind, AIS targets, and hundreds of other data points.

So why is the captain still the one doing all the translating?

Your boat knows:

Had a pretty painful moment last week when one of our support agents got stuck calling a dead endpoint over and over. It just kept spinning and burning tokens in the background until I finally noticed the terminal.

We fixed it simply by capping the retry limit to two in our workflow config and routing it to a human if it fails twice. Cut our runaway token costs right away.

How do you usually handle safety limits or failure thresholds for your agents so they don't get stuck on broken tasks?

One thing I didn't expect while building d3kOS was that the hardest challenge wasn't marine technology, AI, or even running local models on a Raspberry Pi.

It was creating a development process that kept AI from drifting.

After countless experiments, restores, and real-world testing on the water, I ended up creating an AI-focused SDLC and multi-LLM workflow that looks more like a software team than a single AI assistant.