Embed Notice

While I'm doing some thought leadering...

Threat modeling should include the users as threats to themselves.

"How could the most naive person I know use this wrong and suffer?"

or

"How will people actually use this, and does that violate our security assumptions?"

Note that I do not say "stupid" here.

The guy who used sha256("how much would could a woodchuck chuck if a woodchuck could chuck wood") as a Bitcoin private key to protect 250BTC circa 2013 was (is?) a Mensa member.