On Microwaves, Useless Knowledge, and Asking Better Questions
In the article The Usefulness Of Useless Knowledge 1 Abraham Flexner, The Usefulness of Useless Knowledge, Harper's Magazine, October 1939. PDF Abraham Flexner argued that the pursuit of useless knowledge leads to the most important discoveries. His examples were giants in physics. Mine is about buying a microwave. The mechanism is still the same: something you learn without purpose becomes exactly what you need, years later.
The useless knowledge
At some point I watched a Technology Connections video about a Sharp microwave from the 90s. Peak microwave, apparently. The key detail: a humidity sensor that detects steam from the food and stops when it's done. No timers, no guessing, no babysitting. Filed it under interesting and mostly forgot about it.
Recently I was in the market to buy my first proper microwave. Every one before had been a hand-me-down or a gift. I did remember: I wanted one with a moisture sensor.
Teach me the category
This is where GenAI comes in. I've developed a loop for product research that I keep coming back to:
- Teach me the category. What are the dimensions that distinguish a good microwave from a bad one from an interesting one? What technologies exist? Inverter vs. pulsed magnetron, turntable vs. flatbed, ceramic coatings, sensor types.
- Define my dimensions. I care about reheating and defrosting with minimal intervention. I don't care about convection or grill features.
- Find products. Given these dimensions, what's available? Put them in a table.
- Tell me a story. Compare two finalists through a specific use case — reheating yesterday's lunch. Walk me through each, step by step.
That last one is surprisingly effective. Tables show features; stories show experience.
The wish list: sensor, inverter, flatbed, ceramic coating. No single model had all four. The perfect microwave did not exist.
The Panasonic with its sensor: put food in, press reheat twice, wait, done. The LG without: guess a time, check, adjust, guess again. The story made the tradeoff obvious in a way the spec sheet didn't.
The point
I ended up with a Panasonic NN-ST45. Sensor, inverter, simple controls. Not perfect on paper. In practice: I press a button and get well-reheated food. As good as that old Technology Connections video promised.
The AI was useful because I already knew what to ask. That "useless" memory of humidity sensors from a YouTube rabbit hole gave me the one dimension that mattered most. Without it I would probably have optimized for power, easy to clean, maybe UX.
The Usefulness Of Useless Knowledge is about curiosity without an agenda. This is the mundane household version: that odd YouTube video, years later, helped me ask better questions.