We all have human frailties - there's no such thing as a 'dumb mistake'

Watch people and ask questions

It's often more productive to take more notice of what people do than what they say they do. For that you need to watch people using your designs doing real tasks.

To achieve this, we use a process called "user-centred design". We show high-level paper versions of designs to typical users (not people on the design or technical team) in design sessions.
We give them real life tasks to do in a realistic context. ("It's the election. You want to vote for Al Gore, your kids are in the car outside, you have dinner on, and it's freezing in the voting booth.")
Then we watch what they do and ask questions. This uncovers most problems. Then we refine the designs and add more detail, then go back again to discover any other flaws. Finally, we do individual sessions to follow people through a full task, to make certain it's working the way they want it to.

We do all of this before anyone writes a line of code. That's because it's way cheaper to get designs right on paper than it is to rewrite code - about 10 per cent of the cost of changing it at the "prototype testing" phase, and a tiny 1 per cent of the cost of doing it after the system is released.

Get real users involved during the design process and it's much easier to help humans avoid making dumb mistakes with clever products.