[Cryptography] Some quantum computers might need more power than supercomputers

[Michael Kjörling]

On 12 Jan 2026 07:10 -0800, from kentborg at borg.org (Kent Borg):
>> Or an encrypted document stored in one place, and the password for it
>> in another; alongside with some indication of its contents and
>> non-technical instructions for opening it in readable form. It doesn't
>> even need to contain things like passwords itself; it just needs to
>> give one's heirs a starting point for untangling the ball of yarn.
> 
> Yes.
> 
> But it is tricky. One needs to regularly backup passwords, in a way that is
> recoverable by a one or several specific third party(ies), while not being
> recoverable by any other third parties.
> 
> Easy to get wrong.

True; but what I noted is orthogonal to backups. Backing up, say, the
password database is a way to ensure that access to that encrypted
password database is not lost; but it has no bearing on accessing the
_contents_ of the database (assuming access to the ciphertext) as
readable plain text. Having a defined way by which a specific other
party can gain access to the _contents_ of a password database that
this party _has_ access to is a related, but separate, problem.

One possibility which does not rely on any specific vendor support
would be to (n,m) secret-split the key to an encrypted document which
in turn contains the current database passphrase, or recovery details
and information on where that encrypted document is available; then
distribute those shares to trusted people together with instructions
for how to reconstitute the full secret in "open only in the event of
the death of J. Doe" sealed envelopes. The key to that encrypted
document (the key which does not by itself provide access to the
password vault) can relatively easily be rotated.

-- 
Michael Kjörling
🔗 https://michael.kjorling.se