Why this series exists.

Bitcoin has a quantum problem that isn't the problem everyone thinks it is. The media is broken. The experts are cagey. The people actually building the defenses publish in IACR preprints and Delving Bitcoin threads where nobody looks. So I went and read everything.

Who I am

I'm NVK — CEO of Coinkite, the company behind Coldcard, the hardware wallet used by Bitcoin holders who care about operational security. I've been in Bitcoin since the early days, building tools that let people hold their own keys. That's the lens I bring to this work: if you're going to secure serious value over decades, you should understand the actual threat model, not the one being sold by VCs raising rounds on fear.

Why I wrote this

In early 2026 I kept seeing headlines about quantum computers breaking Bitcoin — every few weeks, always citing the same three experts, always with the same doomsday framing. I'd read the underlying papers and the articles didn't match. The industry selling quantum computers has a staggering financial incentive to keep the threat alive. The media amplifies without checking. And Bitcoin developers who know the real state of things get drowned out because they publish in technical venues instead of press releases.

So I went and read everything — every academic paper, every Delving Bitcoin thread, every Bitcoin Optech newsletter, every mailing list argument, every testnet report. Then I wrote it down. Three parts, roughly 15,000 words. Every claim is sourced. (Part 1's title says “200 hours.” The number is rhetorical. The research isn't.)

Methodology

Each article is grounded in primary sources: IACR preprints, Delving Bitcoin discussions, BIP specifications, Bitcoin Optech newsletters (issues #307 through #399), academic papers on quantum resource estimates, and the actual code and testnet reports from the people building this stuff. For Part 3 alone, the compiled knowledge base references 95+ individual research documents.

None of this research lives in a pile of open tabs or scratched notes. It lives in llm-wiki — a personal, LLM-compiled knowledge base I built to make research like this tractable. The pipeline is simple:

1. Ingest — raw sources (papers, blog posts, mailing list threads, testnet reports) get dropped into the wiki verbatim, with frontmatter capturing title, URL, type, tags, and a one-line summary.
2. Compile — Claude reads the raw sources and synthesizes them into cross-referenced wiki articles: concepts, people, proposals, timelines. Every claim carries a confidence level. Every article links back to the sources that support it.
3. Query — I ask questions against the compiled wiki to find gaps, contradictions, and connections I wouldn't have spotted reading linearly.
4. Output — the articles you're reading on this site are themselves outputs of the wiki, synthesized from the compiled knowledge and versioned as new developments land (like Avihu Levy's Quantum Safe Bitcoin proposal on April 9, 2026, or Olaoluwa Osuntokun's zk-STARK BIP-32 escape hatch).

Agentic Research

The research isn't one person reading one paper at a time. It's parallel agentic research: I dispatch a swarm of Claude agents, each assigned a specific angle — academic, technical, applied, news, contrarian, historical, adjacent, quantitative. Each agent runs its own searches, evaluates findings against a credibility rubric, and returns ranked sources. The results get deduplicated and merged before anything lands in the wiki.

For a topic like post-quantum Bitcoin — where the relevant work is scattered across Delving Bitcoin, IACR ePrint, arXiv, Bitcoin Optech, GitHub, mailing lists, and X threads — this is the only way to get comprehensive coverage in a reasonable amount of time. A single researcher reading in order would miss connections between sources that different agents find simultaneously. Eight agents running in parallel, each going deep on one angle, covers in an hour what would take a solo reader a week.

It's not automation-of-judgment. Every article is still hand-edited, every claim is verified against its primary source, every conclusion is mine. The agents are for discovery and synthesis — not for writing conclusions. The difference between “I have an opinion about X” and “I have an opinion about X and here are 95 cross-referenced sources that support it” is the difference this tooling makes. The research is the part you can't see. These three articles are what falls out the other end.

What this series is not

It's not a dismissal of the quantum threat. Bitcoin's cryptographic assumptions are load-bearing and should be diversified regardless of whether large-scale quantum computers ever exist. A $2 trillion network shouldn't depend on a single hardness assumption.

It's also not a panic piece. The researchers building Bitcoin's post-quantum defenses are real, and they're doing serious work. The actual bottleneck is governance, not cryptography. And governance is Bitcoin's thing — we'll figure it out, the way we always do, too slowly and with too much arguing, but we'll figure it out.

Links

The one-sentence version

The house isn't on fire, the builders are working, one of them just demonstrated a door that fits most apartments, and another proved you can barricade your own door without waiting for building management — it just costs more.

— NVK