Skip to main content

The mechanism, examined

A dead man's switch is a good idea with one bad assumption

The assumption is that silence means death. It usually does not — and the release it triggers is the one kind of mistake you cannot undo.

A dead man's switch does something reasonable: it watches for activity, and when activity stops for long enough, it hands over what you left behind. The mechanism is old, it is simple, and it needs nobody's permission — which is exactly why it appeals.

The problem is what it can actually observe. It can observe that you did not open an email. From that single fact it has to decide whether to do something permanent.

Where it breaks

Four failures, all of them ordinary

It cannot tell the difference between silence and death

A timer measures exactly one thing: that nothing happened. A long hospital stay, a stolen phone, a country with no roaming, a hard month, a deliberate break — all of them look identical to a countdown, and all of them can end with your secrets sent to somebody.

The failure is not recoverable

A wrongly-delayed release is an inconvenience. A wrongly-executed one cannot be taken back: a seed phrase that has been seen has been seen, and a private letter that has been read has been read. The two errors are not the same size, so treating them symmetrically is a design mistake.

It makes you responsible for staying alive on a schedule

The mechanism only works if you keep proving you exist. Miss enough emails and it fires; disable it because you are travelling and it protects nothing. Both of those are common, and both are the user losing an argument with their own tooling.

It ends at the moment the family's problem begins

A switch delivers keys. It does not tell anyone which bank, which policy, what to claim, in what order, or with which form. Handing a grieving family a set of credentials is not the same as helping them.

What LifeWyn requires instead

Seven stages. Only the last one discloses anything.

Inactivity still starts the process — it is a genuinely useful signal. It just is not allowed to finish it.

  1. Normal

    Discloses nothing

    You use the app, unlock with biometrics, confirm a periodic check-in. Nothing is pending and nothing is watching in a way you have to think about.

  2. A check-in is missed

    Discloses nothing

    Nothing happens except a reminder to you, on the channels you chose. This resolves the overwhelming majority of cases, by you opening the app.

  3. Reminders escalate

    Discloses nothing

    More attempts, more channels, a longer window. Still nobody else is contacted. Still nothing is disclosed.

  4. Trusted contacts are asked

    Discloses nothing

    The people you designated are asked whether you are alright. They are not shown your vault. Their answer alone changes nothing — it is a signal, not a verdict.

  5. Verification opens a case

    Discloses nothing

    Evidence is gathered and scored, including fraud checks. Only a genuine multi-signal result records the event, and it is cross-checked from a separate service rather than believed on sight. Ambiguity resolves to no.

  6. Any delay you set has to elapse

    Discloses nothing

    A time lock is a per-item, per-recipient condition you configure: this item opens to this person no earlier than this moment. Where you set one, a verified result still waits it out — which is how you keep a wrong answer reversible while it costs nothing.

  7. Your rules are evaluated

    Discloses

    Per item, per person, with any delays and conditions you set. Key shares are decrypted on each recipient's own device — the server never reconstructs your key.

Every gate in that chain fails closed: when a check cannot be completed, the answer is no. That is a deliberate asymmetry — a delayed release is recoverable and a wrong one is not.

Questions

The ones people actually ask

Is LifeWyn a dead man's switch?
It contains one and does not stop there. Inactivity detection is the first stage of the chain — the trigger that starts a conversation. What separates LifeWyn is that inactivity is never sufficient on its own: corroboration and scored verification sit between silence and disclosure, and every one of those gates fails closed.
What if I am simply travelling or off-grid?
Nothing happens except reminders, which you clear by opening the app. Check-in intervals are yours to configure, and periods where you expect to be unreachable can be accounted for in advance. At no point does an unanswered notification disclose anything.
What if I am unconscious in hospital for weeks?
That is the case a plain switch handles worst, and it is why hospitalisation and incapacity are separate states here with their own rules. Someone can be given medical directives and emergency contacts without any part of your estate being opened. Death is not the only condition the system understands.
Could someone trigger it deliberately to get at my data?
The chain does not start with anyone else's assertion — it starts with your own missed check-ins, and you end it by logging in. Beyond that, verification is scored rather than asserted: a scripted burst of submissions, an unverified or tamper-flagged document, or more than one active case all raise fraud risk against the confidence score, and a single contact saying you are fine caps it outright. The release gate is then cross-checked by a service that does not take the event on trust, and every step is audited.
Is a plain switch ever the right choice?
Yes. If your only concern is that a seed phrase should not die with you, and you accept the false-positive risk, a simple switch is cheap, predictable and needs nobody's cooperation. It is a real trade-off and we would rather say so than pretend there is only one sensible design.

Set the conditions yourself

Check-in intervals, channels, who is asked, what delay applies to which item, and who receives which item. All of it is yours to configure.