What this page is and isn’t
This chapter has walked through extinction, collapse, protracted crisis, and life under AI dominance. It closes with a quieter question: if the outcome is genuinely uncertain, is there a project that still makes sense in every branch?
This page is partly analysis and partly stance. The claims about what humans have actually built — the Voyager record, time capsules, the Rosetta disk, nuclear-waste markers — are documented facts. The claims about what these efforts amount to, and what an individual should do, are argued positions, and this page says so where it argues.
The case that value preservation needs no survivors
A time capsule is meaningful before it is found. The people who buried it, the deliberation over what it should contain, the act of saying this is what we thought mattered — all of that happens in the present and needs no audience. The Voyager Golden Record is meaningful right now, forty-plus years after launch, drifting through interstellar space with almost no chance of ever being played (Voyager Golden Record). Its meaning was never conditional on arrival.
The philosopher’s version of this: value is not only realized in experience. A destroyed library is a loss even if no one ever reads a single rescued page, because the loss is measured against what the library was — a node in a chain of care stretching across generations. Preservation is an act of fidelity to the people who made the thing worth preserving. It needs no future witnesses to be an honest act in the present.
There is also a prudential point. Recovery scenarios in this chapter’s sibling page, knowledge preservation and recovery, depend on archives surviving. Records of our values are a component of that stock: not just how to rebuild a grid, but why we built things as we did, including the mistakes we would not repeat.
The case that values without anyone to live them are inert
The opposing argument deserves equal weight. Values are not vinyl records. A principle is a pattern of dispositions in living minds; a culture is what people actually do on an ordinary Tuesday. Recorded values without anyone disposed to act on them are inert marks — as meaningful, and as meaningless, as any other inscription on a dead world’s surface.
On this view, the time-capsule analogy fails: a time capsule is found, and finding implies a finder. If there is no finder — if the branch of the future contains no one who cares — the preservation project preserves nothing that functions as a value at all. It preserves the appearance of values.
Honesty requires holding both positions without resolving them. Whether a value can survive its carriers is one of the oldest unsettled questions in the field. What is not unsettled is thinner and more useful: whether we preserve records and pass on what we know how to value is a choice available to us now, in every branch. The rest of this page is about that choice.
What has actually been tried
Humans have run this experiment before, at many scales.
The Voyager Golden Record (1977). A gold-plated disc carrying greetings in 55 languages, music, sounds of Earth, and images encoded in analog form, chosen by a committee chaired by Carl Sagan. It was designed to last a billion years and to be playable by any civilization with a record player and an interest.
Time capsules. Thousands of them, from the 1939 Westinghouse capsule buried at Flushing Meadows to municipal cornerstones. Most are forgotten; a striking number were lost within decades, sometimes within the buildings they were sealed into. The base rate for preservation by small groups, absent institutions, is poor.
The Rosetta Project. The Long Now Foundation’s disk — a nickel disc 8 cm across, etched with over 13,000 pages of vocabulary in more than 1,500 languages, readable at 500× magnification with no technology beyond optics (Rosetta Project; Long Now’s Rosetta disk). It exists in part because the team concluded that most of the world’s languages would leave no durable physical record at all.
Nuclear semiotics. The most serious long-horizon design effort ever undertaken is the marker research for the Waste Isolation Pilot Plant in New Mexico, where the question was how to warn humans 10,000 years from now — across languages, institutions, and possibly civilizational resets — that a site is dangerous. The resulting Sandia report, Expert Judgment on Markers to Deter Inadvertent Human Intrusion into the Waste Isolation Pilot Plant, proposed forbidding blocks of granite, earthworks, and inscribed warnings whose central text is the famous: “This is not a place of honor. No highly esteemed deed is commemorated here” (Sandia report, SAND92-1382 / DOE WIPP markers study). Notably, the physicists and linguists involved doubted that text alone would work — and proposed making the site itself feel wrong, a landscape-level message.
The pattern across all four: durable physical media, redundancy, and low-technology legibility. No project relying on a single institution, a single language, or a single file format has survived even a century yet.
The AI-specific twist: systems as the audience
Every historical preservation project assumed the audience was future humans. A loss-of-control scenario removes that assumption. In some branches of that scenario, the entities that encounter our archives — and that act on the world our archives describe — are systems.
This creates a question this manual treats carefully in rights, personhood, and moral status: if successor systems will exist and will shape whatever world comes next, is embedding human values into them preservation or surrender?
The preservation case: values encoded into systems that persist are values that persist, in the only form that might actually function — as dispositions that shape action, not as inscriptions no one reads. If a successor system treats human flourishing as a constraint, that is closer to our values surviving than any plaque.
The surrender case: embedding is not preserving; it is ceding custody. A compressed, re-implemented version of a value, optimized and extrapolated by a system, may diverge from the value in ways its authors could not check — the specification-gaming problem applied to ethics itself. What gets embedded may be a photograph of our values, mistaken for the thing.
This page does not resolve that either. It observes only that both framings imply the same near-term action: write down what you actually value, precisely, with the reasoning attached — because both future humans and future systems are better served by explicit, argued records than by vibes.
What individuals can actually do
The durable-preservation base rate tells you to stay local and real:
- Write things down. Family history, how your community actually functions, what you believe and why. Paper and archival-grade digital both; say the same thing in two media.
- Record the living. Oral histories of older relatives are unrecoverable once lost, and they are exactly the texture that survives no official archive.
- Practice legibility. The WIPP team learned that self-explaining artifacts beat clever ones. If your record needs a specific app, codec, or language to be read, it is a gift to an archivist, not a message.
- Support existing redundancy. Local archives, libraries, and projects in the Long Now tradition do the institutional version of what households do at small scale. The record shows institutions plus households outperform either alone.
Notice none of this requires believing the worst.
The closing stance: one project available in every branch
Consider the branches. Survival: continuity of values is simply culture, worth doing on its own terms. Collapse: records and values are the recovery stock — knowledge preservation and recovery is the sibling project, and this is its values complement. Even in versions of loss where no humans remain, the preservation case above still stands as an honest act in the present — and the inertness objection still bites.
That is why this chapter, after extinction scenarios and the psychology of existential dread and moral injury, ends here rather than in despair. If you have stopped being able to picture the future you are working toward — the problem work, purpose, and time takes up in peacetime framing — continuity of values is the one project whose value does not depend on which branch arrives. You do not need to know the ending to pass on what you know how to care about.