AI
Running AI on your own hardware: when a local model makes sense
We run open models on our own machines next to Claude. What we measured, which jobs a local model handles well, and why the harness around it mattered more than the hardware.
4 min
AI
Context is the cost: what we learned measuring our own AI token use
We measured where the tokens go in our own AI-assisted work. The answer wasn't the instructions or the output. It was everything the AI kept carrying around. Here is what we changed.
4 min
AI
Prompt caching for owners: why the second question costs less
If your AI workflow sends the same long instructions every time, you may be paying full price to re-read them. How prompt caching works, what it saves, and when it doesn't help.
4 min
AI
Haiku, Sonnet or Opus: route AI work by how much judgment it needs
Most small-business AI bills are high for one reason: every job goes to the biggest model. Here is how I decide what goes to a small, middle or large model.
5 min
Other tags
Can AI help?
What's the task your team does manually every day?
Tell us in plain words. We'll ask a couple of questions, then tell you honestly whether it's worth automating — no sales pitch.
We've got a clear picture — and a few ways to automate it.
Drop your details and our team will follow up to walk through the options and what it'd take to save you that time. We'll send a copy to your inbox too.