In 1981, Brent Smith wrote a short story about a computer that becomes aware of itself. It is called Room 5000, and when Brent finally published it in 2024, it went out with, in Brent's words, "not one word added, deleted, or moved."
Forty-five years later, the machines are catching up with the story, and Brent is building the locks for what comes next.
What they are building
QuantumViva is a security company in Dublin. Brent describes the vision without any jargon: building lighthouses in a storm. The storm is the fast-growing flood of internet and AI-driven attacks. The lighthouses are places where people can see each other clearly and keep their files, mail and messages private inside it. "Our moral core is personal privacy," Brent says, and admits that is an unusual thing for a company to say.
Its first product, QVUTM, is a single app for Windows and Mac, with Linux and larger companies to follow. Today, protecting sensitive work usually means a patchwork of four to six separate tools: a password manager, an encrypted mail service, a file transfer tool, and often nothing watching the computer itself. Every tool is another account, another key, and another gap. QVUTM puts a secure vault, private email over the accounts people already use, direct device-to-device file transfer and monitoring of the computer into one place. It installs in about ten minutes. Users hold their own keys, so no cloud provider, outsider or government can read their data.
Post-quantum protection is only part of it. QuantumViva also offers:
- a signed evidence record that gives a forensic trail;
- a secure vault;
- no phone number needed to sign up;
- no readable content on QuantumViva's servers;
- the option for a firm to pay for its clients' access.
The roadmap includes versions for larger organisations, digital identity, an in-app security and incident monitoring module built on machine learning, and a digital recovery capsule.
The company has spent two years designing before writing production code: a software specification of more than 500 pages, a full threat model, its own protocol specifications and a complete interface design. This is not Brent's first company. Brent founded Clearwire Ireland and grew it from zero to 26,000 customers and 100 staff on €52 million of foreign investment, and also founded Ardan Audio, which designed audio equipment sold on four continents.
Why quantum computers are a problem for today's secrets
Almost everything that keeps data private today, from banking and email to company networks and the cloud, relies on maths problems that ordinary computers cannot solve in any reasonable time. A large enough quantum computer could solve them. Nobody knows exactly when that machine will exist, but the expected date keeps moving closer.
The urgent part is something called "harvest now, decrypt later." Attackers can copy encrypted data today and simply wait until they can read it. So anything that must stay secret for ten years or more, such as medical records, legal files, inventions or customer data, may already be at risk.
The fix is a new kind of encryption, known as post-quantum cryptography, that even a quantum computer cannot break. The US standards body NIST published the first of these standards in 2024, and the deadlines are now real: US federal systems face 2030 and 2031, EU countries are expected to start moving by the end of 2026, and the UK expects organisations to finish by 2035. Brent says it has similarities to the much smaller but, at the time, monstrous, Y2K upgrade. At its simplest, this is like an upgrade to your existing computer, not the purchase of a new quantum device. They simply don't exist yet as consumer devices.
QVUTM uses a hybrid approach, which European regulators prefer: every message is locked with both a proven classical method and a new post-quantum one, so an attacker would have to break both.
Brent explains the science behind all this for non-experts in a newsletter, The Fragile Sea. Its quantum series, written "in a kind of wonder" to see whether hard ideas can be put in plain English, runs from The quantum sea to The particle zoo and A mansion on the hill, on how quantum computing differs from the ordinary kind.
Others are working on the same shift. PQShield sells post-quantum cryptography to chipmakers and large companies, and SandboxAQ helps big organisations find and replace weak encryption. Signal has added post-quantum protection to its messages, and Proton and Tuta offer encrypted mail and storage. Most of them protect one channel at a time, or serve very large organisations. At launch, QuantumViva is aimed at individual users and small businesses. Each licence covers a Pair, so two people can connect straight away. This is likely to suit professionals and their clients, and small firms that handle confidential work, such as solicitors, accountants, clinics and brokers. It puts secure mail, file storage and signed records in one place on their own computers, and nothing readable is held on QuantumViva's servers.
The point: let AI do the watching, keep people doing the judging
Post-quantum security is brand new, and it changes every week. Draft standards are revised, libraries are patched, researchers publish new attacks, and patents appear. No small team can read all of it by hand.
So Brent set up a weekly "standards watch" using an AI research tool. Every Tuesday at 9am it sends a digest covering the standards bodies, the European certification schemes and the open-source libraries QuantumViva relies on, with a source link for every item. It keeps a standing list of risks it reports on every week, even when nothing has changed, and it says "no notable updates" rather than padding an empty section.
It has already paid for itself. While the encryption design was still being settled, the watch flagged a public patent disclosure at the IETF, the body that sets internet standards. It raised a question about whether the free licence for a new algorithm also covers the hybrid way QuantumViva wants to combine it with classical encryption. That is not a coding problem, it is a "can we ship this at all" problem. Because it surfaced early, getting a legal opinion became a task to finish before the design is locked, instead of a nasty surprise afterwards.
The other half of the rule matters just as much. QuantumViva deliberately does not use AI for threat modelling, the work of deciding who might attack, what they want and which risks to accept. "An AI can produce a threat model that looks complete in seconds. That's the danger," Brent says, "because something that looks complete is easy to stop questioning." Those judgements belong to people who understand the system and will answer for the result. A person signs off every security document and design.
What they are looking for
QuantumViva's protocols have withstood several validation passes with no red flags, even so, it is reaching out for independent validation, and would welcome interested parties. In addition, strategic development partners and additional investment are also welcome, for a quicker path to market, and as always, experts with solid experience in coding secure communications and post-quantum cryptography: please visit quantumviva.com for contact details. QVUTM is coded in c/c++, rust, and a python (pyside6) GUI.
Where else this works
This part is opinion rather than anything the company has said.
The rule fits any small team working in a field where the rules move faster than people can read them. A medical device maker could have AI watch regulators' guidance every week. A fintech could watch new rules in every country it serves. A drug developer could watch patent filings, a food exporter could watch labelling laws, and a battery maker could watch safety standards. In each case the AI does the tireless reading and flags what changed, and a person who will answer for the result decides what it means.
Why it is bigger than it sounds
The short version is a security app. The bigger idea is simple.
- Your secrets can be stolen today and read later. That is why protecting data against quantum computers cannot wait for the computers to arrive.
- One app instead of many. Fewer tools means fewer keys, fewer accounts and fewer gaps for an attacker.
- AI watches, people decide. The machine reads everything every week, and a person makes every call that matters.
What to watch
The measure is independent review. Security people trust what outsiders have tried and failed to break, so an academic partner testing QuantumViva's designs matters more than any claim the company makes itself.
The second is the patent question. A clear legal answer on the hybrid design clears the way to build.
The third is QVUTM's launch, and whether people doing sensitive work will move from a pile of tools to one.