S16 E45: Shai Wyborski on Bitcoin, Kaspa & Proof of Work Research

This turned out to become the longest episode in the history of the Bitcoin Takeover podcast… yet: 8 hours and 39 minutes, featuring Shai Wyborski. He is a co-author of the GHOSTDAG paper and one of the most independent thinkers in proof of work research. Shai dropped out of high school at 14, found his own path into academia, earned a PhD in quantum computing, and then helped design the protocol that would become Kaspa. He also brought his guitar, so if you stick around long enough, you get a jam session between two podcasters who refuse to stop talking.

Shai’s Bitcoin story begins around 2011, and yes, he did have around 1,000 BTC at some point. He spent them, because back then people still talked about Bitcoin as a medium of exchange. What did he buy? “Stuff that’s good for creativity.” He is not at liberty to discuss it further.

His disenchantment came in two blows: first when Steam started accepting Bitcoin and then dropped it months later because congestion made payments unreliable, and then the SegWit saga, which to him felt like tyranny from Bitcoin Core. He is quick to admit this view was naive, and that being closer to protocol development has made him understand the Core side much better. But at the time, it pushed him away.

What brought him back was Daglabs, Yonatan Sompolinsky’s company, which eventually settled on implementing GHOSTDAG and launched Kaspa in November 2021. And this is where the conversation gets interesting, because Shai no longer believes Kaspa makes Bitcoin obsolete, nor does he like the store of value narrative. Instead, he offers a third thesis: Bitcoin as the ultimate decentralized vault, the Fort Knox of data and proof of publishing.

If you want to make a commitment to data that people will still trust in 200 years, you use the most secure chain in the world, and posting data there should be expensive, because high fees are exactly what keeps the vault secure. Meanwhile, fast chains like Kaspa could anchor their state commitments into Bitcoin, letting anyone verify decades of history against the immutable ledger. He calls it maybe the most beautiful application of Bitcoin he can think of: humanity documenting itself in a way that is not censorable or retroactively coercible.

But the meat of the episode is Shai’s criticism of Kaspa itself, and this is what makes him such a rare guest: a co-author of the protocol who is willing to say, on record, that the fair launch model is failing. His argument is subtle. A fair launch and a community project are not the same thing.

A fair launch without a social contract between developers and wealthy community members leaves the project vulnerable to capture: someone with enough money (say, an ASIC manufacturer who mined 80% of the supply for three months before announcing their machines) can simply start footing all the bills, gain an exclusive channel to the developers, and exert disproportionate influence. Shai reveals that three individuals collectively hold around 1.5 billion KAS, roughly 6% of the supply, and that the centralization of community funds has made the ecosystem hostile to outside builders. He knows of developers who tried to build on Kaspa and gave up because documentation was scarce and the devs were too busy to answer questions.

Then there is the governance drama. A few months before our recording, a conflict erupted between Yonatan Sompolinsky and parts of the community over how Kaspa communicates publicly. The result, three months later, is a total blackout: no vision, no plan, no representation. Shai’s verdict is blunt: being a great technology visionary does not make you a great leader, and the reclusiveness of the core team, combined with their dismissal of outside research (including a detailed fee market analysis by a PhD in financial mathematics that core never acknowledged), is centralizing development through social separation.

Shai himself was cut off mid-project from funding for his Kaspa book, which he calls the biggest mistake of his career. The good news: the book lives on as a general proof of work textbook, now funded by an education grant from Quai Network, and you can already read large parts of it online.

On the technical side, this episode is a masterclass. Shai explains what selfish mining actually is (withholding blocks to knock honest blocks off the chain, not mining empty blocks, a mistake I made and was happy to be corrected on), and how Qubic’s attack on Monero changed his mind about its urgency. He estimated Qubic controlled only around 28% of Monero’s hashrate, yet still managed to severely disrupt liveness, precisely because RandomX relies on commodity CPUs, the most abundant and easily rallied hardware in the world.

His advice to Monero is simple: drop RandomX, because the CPU mining experiment has failed the test of reality. His broader claim is one he is willing to die on: the only way to sustain decentralization in proof of work is an ultra ASIC-friendly algorithm, forcing mercenary miners to buy hardware that can do nothing else.

His case for proof of work over proof of stake is equally sharp. Any system where influence is derived from an intrinsic resource (coin holdings, reputation, delegation) codifies rich-get-richer dynamics directly into the protocol. Proof of work tethers influence to external markets, energy and hardware, which are chaotic, competitive, and constantly shifting, and that instability is a feature. His one rule for hybrids: proof of stake can govern anything except transaction inclusion and ordering. The moment PoS touches sequencing, you have a PoS chain with extra steps.

You will also learn why GHOSTDAG is “literally the worst name in crypto” (it generalizes Nakamoto consensus, not GHOST, and Kaspa is simply Aramaic for silver), why DAGKnight might be the perfect proof of work protocol (scalable, strong liveness, and self-scaling, adapting to real network conditions instead of hardcoded assumptions), and my favorite Easter egg ever embedded in a monetary policy: Kaspa’s monthly emission reductions follow the frequencies of the chromatic musical scale, starting at 440 KAS per block, like the middle A of a piano. Kaspa is the only proof of work you can use to tune a piano.

Shai also reveals he mined 5 million KAS in the first days and gave all of it away in paper wallets: to his grocer, his psychologist, his family. At the top, that was a million dollars. He does not regret it.

The quantum computing segment alone is worth the price of admission. Breaking Bitcoin’s ECDSA requires around 2,300 logical qubits; the best machines today have two or three. A survey of leading experts found that more than half believe there is a greater than 50% chance of cryptographically relevant quantum computers by 2035, which for Shai is enough to start preparing now. He dismantles the Y2K comparison made by skeptics (Y2K was a non-event because thousands of engineers worked for years to fix it, just like the ozone layer), explains why post-quantum signatures are 15 to 20 times larger and thousands of times heavier to compute on hardware wallets, and presents his own research on quantum canaries: deliberately weakened UTXOs acting as on-chain tripwires that give the network a measurable, game-theoretically sound signal of when quantum hardware crosses dangerous thresholds.

He also covers quantum mining, which is wildly counterintuitive: rational quantum miners increase orphan rates just by existing, and Grover’s square-root scaling means a miner manipulating difficulty upward on a fast-adjusting chain like Kaspa could theoretically rewrite history with a fraction of the hashrate. The machines required are so absurdly powerful that this remains a curiosity rather than a threat, but it is a fascinating lesson in how many of our security assumptions are artifacts of classical computation.

We also get into why Kaspa’s fee market can activate without congestion (parallel blocks force miners to randomize transaction selection, so paying more always buys you better inclusion odds), why Lightning centralizes into supernodes, why privacy coins face an unwinnable fight against TradFi compliance, why the crypto industry is “selling fundamental science directly to retail” (his analogy: a chemist inventing a new car material and pitching it on the street instead of to manufacturers), and why FUD is the single best educational opportunity a researcher can get.

And because this is the Bitcoin Takeover podcast, we also spent an hour on philosophy (absurdism, utilitarianism, and the categorical imperative are the three legs of his worldview; Ayn Rand “solved” the metaphysical gap by declaring “existence exists”), an hour on music (Sgt. Pepper invented the studio album; Pet Sounds made Paul McCartney jealous; he prefers In Utero to Nevermind), and a memorable stretch on what it means to be Israeli in crypto when strangers demand you answer for geopolitics before they discuss your math.

Shai beat Yonatan’s record by over an hour. If anyone ever breaks his, he promised to come back.

This episode is sponsored by Sideshift.ai, NoOnes.com, Bitcoin.com News, LayerTwo Labs, and Citrea! Special thanks for making this 8.5-hour marathon freely available to everyone!

Listen to Shai Wyborski on YouTube, Spotify, Apple Podcasts & more!

▶ Play video Watch on YouTube

Time stamps:

00:01:09 – Intro: Shai Wyborski, GHOSTDAG & the 8-hour record
00:03:41 – Dropping out of high school at 14
00:04:44 – Shai’s evolving opinions on Bitcoin
00:05:21 – Spending 1,000 BTC on “creativity fuel”
00:06:31 – Steam drops Bitcoin & the SegWit “tyranny”
00:10:47 – Bitcoin as the decentralized Fort Knox of data
00:15:44 – Bitcoin’s problems: throughput & Lightning’s limits
00:16:31 – Kaspa’s multi-leader consensus explained
00:23:24 – The failure of the fair launch model
00:37:42 – Three individuals hold ~6% of KAS supply
00:41:44 – Sponsor reads with live guitar jam
00:52:03 – Working with Quai Network (disclosure)
00:53:06 – Qubic’s attack on Monero with ~28% hashrate
01:08:27 – What selfish mining actually is
01:12:17 – Why proof of work beats proof of stake
01:21:24 – Advice for Monero: abandon RandomX
01:24:47 – PoW/PoS hybrids & Zano
01:36:46 – GHOSTDAG: “the worst name in crypto”
01:49:08 – DAGKnight: the perfect proof of work protocol?
01:59:22 – What Shai would change about Kaspa’s launch
02:07:08 – Mining 5M KAS & giving it all away
02:12:01 – Should the GHOSTDAG authors step back?
02:15:05 – The Kaspa governance drama & core blackout
02:28:40 – The proof of work book & the Quai grant
02:45:11 – Why FUD is the best educational opportunity
02:54:33 – Kaspa storage requirements at Visa scale
03:00:55 – Lightning Network critique: supernodes & trust
03:08:33 – Israel, politics & crypto tribalism
03:23:18 – Privacy, miner censorship & compliance
03:35:09 – How Kaspa splits fees between parallel blocks
03:39:53 – Guitar jam session
03:43:21 – Does Shai believe in God?
03:47:26 – Absurdism, utilitarianism & categorical imperative
04:02:32 – Ayn Rand, objectivism & “existence exists”
04:12:15 – Jordan Peterson & decentralized cults
04:25:29 – Does Shai actually use Kaspa?
04:45:20 – Tokens, vProgs & smart contracts on Kaspa
04:55:11 – Stablecoins are just tokens
05:20:03 – Do two pools control 51% of Kaspa’s hashrate?
05:25:49 – Decentralized oracles vs Chainlink
05:48:45 – Quantum computing explained: logical qubits
06:03:47 – 2,300 logical qubits to break Bitcoin’s ECDSA
06:05:37 – Expert survey: quantum attacks by 2035?
06:08:01 – Post-quantum signatures: size & hardware problems
06:14:12 – Quantum loot: Satoshi’s coins & P2PK
06:29:53 – Quantum canaries: Shai’s on-chain tripwire
06:45:25 – Quantum mining attacks & Grover’s algorithm
07:02:29 – Unclonable cryptography & quantum money
07:10:21 – Kaspa’s security budget: fees without congestion
07:18:24 – The chromatic emission: tuning a piano with Kaspa
07:32:48 – The hardest part of earning a PhD
07:44:04 – Beatles, Sgt. Pepper & the studio album
08:21:14 – Breaking the 8-hour record
08:26:33 – What Shai is working on now
08:34:21 – Outro jam & final thoughts