FIELD NOTES
Our thoughts on the industry and beyond.
Do you need an NDA?
An NDA protects information you can prove you shared in confidence, not an idea, for as long as the contract and your secrecy measures last. What the cases and statutes say, and where NDAs fail.
read →EV vs gas: the road runs 10 years behind the showroom
A quarter of the world's new cars in 2025 were plug-ins. The road follows about 10 years later: in 2046, about 4 in 10 cars still run on fuel alone. A fleet model worked from a lot of 100 cars to four regions, with ranges and signposts.
read →Humanoid robots in factories
Even at full human speed, a leased humanoid on two shifts costs about $32.50 per hour of work, a little under the CNC operator it would replace. A six-mill job shop worked through with the numbers: the yardstick, the humanoid in its perfect form, what keeps robots out of small shops, and the machine the night shift needs.
read →How custom sports equipment is developed
Six steps take a piece of sports equipment from a sketch to a product in a box. A lacrosse ball collector worked through each one, with the numbers: the ball, the concepts, a wire-gap test, fatigue, and the quantity at which a mold pays for itself.
read →Custom bioreactor design, what it takes and what it costs
A custom bioreactor is worth building when your protocol needs a vessel geometry, flow, or sensing that no catalog unit provides. Here is what goes into one, what the parts and the engineering cost, and when to modify a catalog unit instead.
read →How to turn an idea into a product
Turning an idea into a product is a sequence of small, decisive steps, not one big build. Here is the honest path, what each step decides, what it costs, and how to hire for it without getting burned.
read →How to spot an invention-promotion scam
You have an idea and some money, you search for help, and the first companies to find you sell patents and marketing, not products. Here is how to tell an invention-promotion firm from a real engineering one before you spend a dollar.
read →When a hardware project stalls, how to restart it
Developer disappeared, prototype half-works, or you inherited an undocumented rig? A stalled hardware project can usually be restarted. The first step is an honest diagnosis, not more building. Here is what a rescue involves and how to judge whether yours is worth finishing.
read →Custom lab equipment without the instrument-company markup
Your experiment needs an instrument no catalog sells. The big vendors quote five figures, and the student who built the last rig has graduated. Here is a better path, and the purchase-order lane your department can approve on its own.
read →What a hardware prototype actually costs in 2026
Quotes for the same idea range from $3,000 to $300,000, and almost nobody publishes numbers. Here are the real ranges, what drives them, the three ways people overpay, and when you shouldn't build at all.
read →