Blog
Notes on data, AI and products that work in the field
Practical writing for founders building AI for machines, vehicles and energy. Data first, then AI.
>>
What Data to Collect Before You Build Physical AI
Most physical AI startups collect too little of the right signal and too much of the wrong one. Here's how to decide what to capture from your machines, vehicles or sensors before you train a single model.
September 29, 2026
·
7 min read
>>
Why Your Physical AI Demo Works in the Lab and Fails in the Field
The demo worked on your test data. On a real customer's machines, it falls apart. It's rarely the model's fault — it's the gap between lab data and field data. Here are the five gaps we see most, and how to close them.
September 15, 2026
·
7 min read
>>
Edge or Cloud? A Founder's Guide for Physical AI Products
Every physical AI startup hits the same question: should the AI run on the device or in the cloud? The honest answer is usually 'both' — here's a plain-language way to decide what goes where.
September 1, 2026
·
6 min read
>>
Do You Need a CDO, a CTO, or Neither — Yet?
Non-technical founders building physical AI often hire the wrong senior role first — or hire too early. Here's how to tell whether you need a Head of Data & AI, a Head of Engineering, a Head of Product, or none of them full-time yet.
August 18, 2026
·
6 min read
>>
How to Scope a Predictive Maintenance Pilot That Converts to Paid
Most physical AI pilots end with a polite 'interesting, let's stay in touch'. The ones that convert are scoped differently from day one. Here's how to design a pilot your customer can say yes to.
August 4, 2026
·
7 min read
>>
What Running 1M+ Connected Endpoints Teaches You About Startup Data
Our founder built a data platform with more than a million endpoints connected at once. Startups are far smaller — but the same lessons apply from the very first device. Here are six of them.
July 21, 2026
·
6 min read