Quote from: JayJuanGee on Today at 06:26:07 PM

Quote from: nullius on Today at 05:32:33 PM

Security and availability is another one of those issues that I intended to write about, on the subject of using seed phrases with ostensibly strong passphrases. Re AlcoHoDL, et al. Have not yet gotten to it. In a nutshell: If you care about your heirs, don’t use passphrases! Use Shamir’s Secret Sharing to back up secret pieces in some way that mitigates risk of compromise while you are alive, but also has a high probability of surviving your death or otherwise incapacitation.

Well, whenever you are writing further on this particular topic (perhaps in the form of a tome or treatise? - #nohomo), hopefully you keep in mind practicality of regular joes - reminds me a bit of how trace mayer used to recommend people hold their keys.. ... Yeah, right, works for technical geniuses, but not necessarily for regular joes.

In other words, not everyone is ready, willing and able to learn, even if they might meet part of that formula.. they need all of it... including the action and even time components.

Not everyone is as technical as you, and the fact of the matter remains that the vast majority of regular joes (and the jane or two that is in this space) need to have simple as fuck.. and prefer to have simple as fuck.

As a practical matter, the hardest part isn’t technical. As usual, the biggest problem is human. A very rough sketch:

How astute are you at judging character? Could you choose N people in your life, such that it is very unlikely that M of them would conspire against you? —And also very unlikely that (N - M) + 1 of them would inopportunely die, disappear, lose stuff, or just flake out?

(Remember that with M/N Secret Sharing, M - 1 shares together reveal zero information about the secret; on a technical level, it is information-theoretically secure.)

Any potential traitors face the problem that to betray, they need to risk potentially suggesting betrayal to somebody who may be loyal—somebody who would promptly inform you. (Though if your threat model may include a risk of coercion of your fiduciaries, this calculation could be turned upside-down.)

Just for a contrived off-the-cuff example: Maybe you don’t (and shouldn’t) trust your lawyer. But how likely is it that he, a member of your family, and your best friend would all form a conspiratorial meeting of the minds for the purpose of stabbing you in the back and ripping you off? —Especially if none of them is aware of what he has? Whom do you know who would safeguard for you a sealed envelope, with instructions to open it if and only if you are dead or permanently incapacitated? (Or return it to you if you ask for it in person—as your last-resort catastrophe-plan backup.)

If you can figure out the human problem, then the technical part is easy!

Of course, all that human stuff is what is usually forgotten by technology people.