The unbroken chain: why digital erasure became cheap, and what we build against it

01The architecture of forgetting

Paper archives were slow, expensive and imperfect, but they had a property modern systems quietly abandoned: physical multiplicity. Copies of anything important lived in libraries, government depositories, newspaper morgues, private collections and personal filing cabinets. Destroying every instance required coordinated, visible effort across many locations. Even regimes that attempted systematic erasure usually left residual traces, and the attempt itself left witnesses.

Digital records inverted the economics. When the authoritative version of a document, a news story, a dataset or a legal filing lives on servers controlled by one institution or a small set of platforms, the cost of erasure collapses. A policy change, a legal threat, an administrative decision or a routine database purge is enough. The URL returns "page not found." Search indexes drop the reference. Within a short time the public record treats the material as if it never existed. Assange observed this happening in real time with major newspapers that quietly unpublished stories under pressure: the old addresses simply stopped resolving, with no acknowledgment that anything had been removed.

Assange tied the phenomenon to Orwell's most durable line: he who controls the present controls the past, and he who controls the past controls the future. In a world where history exists primarily as bits on machines, control of those machines becomes control of collective memory. The deeper problem is not that pages disappear. It is that the disappearance can be silent. A burned library leaves ash and witnesses. A digital purge can leave only the absence of evidence, and the absence of evidence is then offered as proof that the evidence never existed.

02The measurable scale of disappearance

None of this is hypothetical, and it does not require a villain. The Pew Research Center studied a decade of web content and found that 38 percent of pages that existed in 2013 were no longer accessible by 2023. A quarter of all pages that existed at some point between 2013 and 2023 were already gone when the study ran. Harvard and Columbia researchers who examined links inside New York Times journalism found that a quarter of all deep links in articles were already dead, rising past half for older material, and that is one of the best-resourced archives in publishing.

Link rot infects citations in judicial opinions, academic papers and government reports. Entire publications have vanished when ownership changed or companies failed. Platforms remove content at volume, sometimes before independent researchers can preserve a copy. Government websites have undergone systematic removals and rewrites covering public-health data, scientific contributions and demographic records. The Wayback Machine and similar archives are heroic but partial: they are incomplete, they can themselves be pressured, and a snapshot does not carry the original publisher's attestation that the material was authentic at the time.

03Beyond deletion: the quantum threat to authenticity

Deletion is the obvious failure mode. The subtler one arrives when the cryptography that proves authenticity stops being trustworthy. Modern digital signatures, certificate systems and blockchain transaction authorizations rest on two mathematical problems, integer factorization and elliptic-curve discrete logarithms, that classical computers cannot solve efficiently. A sufficiently large fault-tolerant quantum computer running Shor's algorithm changes the economics of both. Private keys become derivable from public keys. Signatures that verify perfectly can be forgeries.

Notice what breaks and what does not. In a blockchain, the chain of hashes linking successive blocks remains mathematically intact; an attacker does not silently rewrite old blocks. What breaks is signature trust: the ability to know that a transfer, an attestation or a state update genuinely originated from the claimed party. Once signatures can be forged at scale, the practical integrity of the record collapses even though the hash chain is continuous. Ownership can be diverted. Attestations can be fabricated. The line between authentic history and manufactured history erodes, and it erodes without a single page returning "not found."

This is not speculative fiction, and the timeline is not ours. NIST finalized the post-quantum standards in August 2024: ML-KEM (FIPS 203) for key exchange, ML-DSA (FIPS 204) and SLH-DSA (FIPS 205) for signatures. National security agencies have published migration timelines clustering around the late 2020s through the mid-2030s. Hybrid post-quantum key exchange already protects a large and growing share of internet traffic at major content networks. The migration is underway because "harvest now, decrypt later" is understood: material protected today under classical cryptography may still need to be confidential or provable years after a cryptographically relevant quantum computer arrives. For records, the equivalent is sign now, forge later: a signature made today with RSA or ECDSA is a promise that expires on a date nobody can name.

04What an unbroken chain actually requires

An immutable record is only as strong as the cryptography binding each link, and the assumptions under which that cryptography stays secure. Elliptic-curve and RSA signatures were designed for a classical world; they were never meant to outlive a cryptographically relevant quantum computer. Relying on them for long-lived records is writing history in a medium a future technology can dissolve.

Post-quantum cryptography replaces those primitives with algorithms believed to resist both classical and quantum attack. When lattice and hash-based signatures are used to sign transactions, attestations and document hashes, the resulting chain inherits resistance to the primary quantum threat against digital authenticity. The practical effect is precise and worth stating carefully:

• Deletion can still happen at the hosting layer. Any operator can stop serving data. No cryptography prevents that.
• What deletion achieves changes completely. The cryptographic commitments remain independently verifiable by anyone holding a copy of the signed material. The record can be re-proved from any surviving copy.
• Forgery becomes computationally infeasible under the new assumptions, so the record cannot be silently rewritten either.

The chain does not silently become editable. It stays unbroken under the threat model that will define the coming decades. That is the whole design goal, and it is a different goal from secrecy: the point is not hiding information, it is durable authenticity, the preserved ability to demonstrate what was published, by whom, and when.

05What KXCO runs in production

This is the problem space KXCO operates in, and the capabilities are live rather than promised. We describe them here at the level of what they guarantee; the recipes stay ours.

The same principles run through the rest of the platform: outbound KXCO email is signed and anchored, records in Meridian carry verifiable attestations, and the ontology keeps identity, authority and events in one structure that every party can check rather than reconcile. The objective is public and narrow: reduce the surface area on which history can be quietly rewritten or forged.

06The stakes: journalism, markets, public memory

Journalism depends on citing sources that remain available and whose authenticity can be checked; when original documents disappear or can be forged after the fact, accountability collapses. Financial markets depend on the integrity of ownership records, transaction histories and regulatory filings. Legal systems depend on signed contracts and evidentiary chains. Science depends on the durability of published data and methods. In each domain the same structural risk appears: centralized digital control combined with quantum-vulnerable cryptography creates the conditions for selective amnesia and selective fabrication.

Any system that makes erasure cheap and authentication brittle will eventually be used by whoever holds power at the moment of decision.

Assange's observation was never only about governments or intelligence services. It was about the architecture of information itself. Paper raised the cost of revisionism through physical multiplicity. Properly designed quantum-resistant systems raise that cost through mathematics that stay hard even for quantum computers. The alternative is a future in which inconvenient pages return "does not exist," inconvenient signatures can be manufactured, and the public is invited to accept the revised narrative because the evidence has been made unavailable or untrustworthy. That outcome is a design choice, not an inevitability.

07Keeping the chain unbroken

The technical path is clear, and it is the same one we apply to our own systems:

1. Migrate signatures and attestations to the standardized post-quantum algorithms, FIPS 203, 204 and 205, starting with anything that must remain valid for years.
2. Design for independent verification. Archival and ledger systems should keep their cryptographic commitments checkable even if primary servers go offline or stop cooperating.
3. Treat long-term authenticity as a first-class requirement, not an afterthought bolted on once a record becomes contested.
4. Measure success on the right horizon: not the absence of today's breaches, but the continued provability of records years after quantum computers become practical.

Assange's warning stays urgent because the incentive to rewrite history never disappears. What changes is the cost of doing it. Paper imposed that cost through physical distribution. Quantum-resistant cryptography and carefully engineered verifiable commitments impose it through mathematics. Societies that understand the difference will keep the ability to know what actually happened. Those that do not will eventually discover that "the page does not exist" was only the first step.

08What we claim, and what we don't

KXCO is a software company operating in the UK and USA. It holds no financial licences and does not custody assets. The institutions that deploy KXCO software operate under their own licences and hold the relationships with their own customers.

09Frequently asked questions

Can a blockchain stop someone deleting a web page?

No, and nothing can. An operator can always stop serving content. What a signed, anchored record changes is what deletion achieves: the cryptographic commitment survives on the chain and in every copy of the signed material, so the content can be re-proved authentic from any surviving copy, and the claim that it never existed stops being available.

Why does quantum computing threaten records that already exist?

Because the signatures that prove those records authentic rest on integer factorization and elliptic-curve discrete logarithms, and Shor's algorithm solves both on a sufficiently large fault-tolerant quantum computer. A record signed today with RSA or ECDSA may need to remain provable for decades, well past the point where such signatures can be forged. That is why the migration matters before any such machine exists.

Does the hash chain of a blockchain break under quantum attack?

The hash links themselves are not the weak point; hash functions are far less affected by quantum algorithms. What breaks is signature trust: the ability to know that a transaction, attestation or state update genuinely came from the claimed party. Once signatures can be forged at scale, the practical integrity of the record collapses even though the hash chain is continuous.

What does KXCO actually use for quantum-resistant records?

The NIST-standardized lattice signature scheme ML-DSA-65 (FIPS 204) for signing, hashes of signed material anchored to Armature L1, a settlement network that has used ML-DSA-65 from its first block, and the open-source kxco-verify package so anyone can check a signature offline with no KXCO server involved.

Do I have to trust KXCO to believe a record is authentic?

No. The algorithms are published NIST standards, the verification tools are open source on npm, and Armature L1 has a public explorer. Verification runs against the signed material and the public record, not against a KXCO service.