S15 E8: Ryan Condron on Bitcoin Mining, Hashrate Markets

Ryan Condron has been mining Bitcoin longer than I have known about Bitcoin. I met him in a hallway in Prague, during what was supposed to be a normal conference chat, and within minutes we were interrupted by a woman from the UN who wanted to know whether Bitcoin mining was causing earthquakes in Turkey.

Ryan did not flinch. He listened. He asked how heat from ASICs was supposed to rearrange tectonic plates. She could not quite land the science, understandably, and I walked away thinking: this guy belongs on a podcast. Patient, technical, and somehow still amused by the circus.

So for S15 E8 I sat him down for about forty-two minutes of mining history, ASIC graveyards, and the odd little idea he has been nursing since 2017: if you cannot stop rich people from buying warehouses full of hashers, maybe you can still stop them from pointing all that hash at the same pool.

That idea is Lumerin. A decentralized marketplace for hash rate, where renting a miner means controlling where its work actually goes, not just collecting a payout spreadsheet.

Garage GPUs, Feathercoin & the Newegg Era

Ryan started at the end of 2012 the way a lot of early miners did: Newegg GPUs, garage benches, extension cords daisy-chained across every circuit in the house. They barely mined Bitcoin itself. Difficulty had already outrun hobbyist graphics cards, so the strategy became mining Litecoin, Feathercoin, and whatever else would still pay, then selling those coins to stack BTC.

“BTC was our big friend at the time,” as he put it. Altcoins were the farm; Bitcoin was the barn.

Solo mining was still the default. Pools were optional until they were not. By late 2013, as more hash pointed at the same niche chains, solo stopped making sense for smaller operators, and Ryan’s crew started building pools of their own. I brought up Slush Pool (originally Bitcoin.cz back in 2010, with Satoshi himself aware of it) as the first real crack in the “leave your laptop open overnight” model. Steady rewards beat lottery tickets. Decentralization took the hit.

Ryan pushed back gently. He is not convinced Satoshi was shocked by pools. Splitting a nonce space across workers is, in his words, a fairly trivial thing. Once Bitcoin is valuable enough to look like a global currency, people will amass CPUs and GPUs either way. Centralization risk arrives with the money, with or without a pool operator in the middle.

From Xilinx Boards to the Algorithm Explosion

The hardware treadmill was already sprinting. CPU to GPU to FPGA to ASIC, compressed into roughly a year of chaos.

Ryan’s cousin worked at Xilinx. In 2013 he messaged her, asked for FPGA development boards, and got high-end hardware courtesy of the company. SHA mining Verilog already existed; he started hacking at scrypt designs to get ahead of the Litecoin-side ASICs that would arrive later that year. Butterfly Labs pre-orders were floating around. Chinese factories were almost certainly already hashing at scale before Western hobbyists got their boxes.

Then came the 2013 algorithm explosion. Every new coin invented its own PoW soup (X11, Cryptonight, Groestl, and eventually something like one algorithm per ticker) in the name of ASIC resistance. Ryan’s favorite cautionary tale is Siacoin: years of “ASIC resistant” marketing, then a sudden in-house ASIC launch. Monetization dressed up as ideology.

He is clear-eyed about what followed. Proof of work’s “everyone mines” vision has fallen out of fashion. New projects ship as proof of stake because it is easier to control, easier to distribute up front, and easier to monetize for the founding team. Long-term, he thinks the concentration risks around those initial token distributors are “incredible.” Short-term, founders get to skip the hardware race.

I made my usual case for PoW as the anti-Ponzi feature: competitive discovery of supply, no central emitter, anyone can join the race. Ryan did not argue against leaving Bitcoin on PoW. He simply mapped how far we have drifted from Satoshi’s interchangeable “node” and “miner,” and how even early Ethereum baked mining into Geth the same way Bitcoin originally bundled wallet, miner, and chain into one app.

Space Heaters, S7 Graveyards & the Six-Month Break-In Joke

Somewhere around 2017–2018 the culture split: Raspberry Pi nodes for validation purity on one side, and on the other a cheerful attempt to turn dead ASICs into living-room heaters. CryptoCloaks and friends commodified Antminer S9 space heaters for roughly two hundred dollars. Mine and warm the apartment at the same time.

Ryan had been living that joke for years. He once ran an S3 in his bedroom through the winter. In the house addition he ran so many S3s that the thermostat hit something like 140 degrees Fahrenheit because he could not vent the heat fast enough. Extension cords to every circuit. The whole early-miner aesthetic.

He remembers Chinese warehouses after the S5 era: mountains of discarded boards thrown into piles because there was nothing else to do with them. Recycling into heaters beats the dump. On durability he is picky. S7s were terrible in his fleet, nearly fifty percent failure within six months. S9s are tanks. Dust-coated boxes still hashing years later. Millions of them basically built the modern Bitcoin mining ecosystem.

Then there is the Newegg trauma. One of the saddest days of his life, he says, was when Newegg started accepting Bitcoin. Spend BTC on GPUs and ASICs, watch difficulty eat your expected coin yield, recover dollars maybe, never the coins. The running joke of that era was that manufacturers had a secret six-month “break-in” period: rack the machines, mine through the easy difficulty, then ship them as brand new once the network had already moved on. Nobody had proof. Everyone told the joke.

Bitmain, Whatsminer & Why China Still Wins

Today’s manufacturing map is thin. Bitmain and Whatsminer (MicroBT) trade blows. Avalon, Canaan, Innosilicon, Spondoolies came and went. Intel went in heavy on SHA-256 silicon and then backed out for reasons Ryan does not claim to know. US teams talk about better architectures; he reminds you that a clever chip is one step in a hundred-step journey to a shippable device.

China wins on cost of manufacture. North America, in his framing, is a terrible place to build consumer electronics at scale. Currency wars and living costs do more to centralize the ASIC supply chain than any particular CEO. He mentioned Norwegian Litecoin-side chip makers we had lunch with in Prague; impressive tech, still swimming upstream against Chinese unit economics.

Even the gear on our call was a quiet punchline. Microphones, cameras, phones: Chinese-made whether the brand sticker says so or not. Western consumers want cheap gadgets. Cheap gadgets need cheap factories. Until Latin America or Africa absorbs more of that manufacturing, the dominance holds.

Decentralization of Control, Not Just Geography

This is where Lumerin enters, and where Ryan stops being a nostalgic miner and starts being a protocol guy.

In 2017 he wrote a paper called Decentralization of Control. The distinction he cares about is distributed versus decentralized. Distributed is geography: miners in Texas, Kazakhstan, Paraguay. Decentralized is who decides where the hash points. You can have a fault-tolerant, globally scattered network that is still controlled by a handful of publicly traded fleets.

“The moment you show someone a little toaster and you tell them you plug it in it’ll print you money, they’re going to quickly say how many can I have and where can I put them.”

Hundreds of millions of dollars later, you get facilities with tens or hundreds of thousands of machines, and public miners sitting on thirty to forty exahash or more. Optimization loves centralization. Public markets love optimization. Ryan’s conclusion is blunt: you are not going to keep the toasters out of rich hands.

So stop fighting ownership. Fight control.

On Lumerin, a Texas warehouse with 100,000 devices does not have to point every board at one corporate pool. The operator lists hash streams on a decentralized marketplace. Buyers anywhere purchase time-bound rights to those streams. The kicker, and this is the part that separates Lumerin from classic cloud mining, is that buyers purchase control of the endpoint, not merely a claim on rewards. They decide which pool, or whether to solo, or how to amass ideological hash. Ownership stays with the facility. Agency moves to the market.

Miners become a utility: hash power producers selling to the highest bidder. Buyers remain anonymous if they want. The commodity is real work, routed over a peer-to-peer system.

Not NiceHash With Extra Steps

I asked the obvious question. Cloud hashing and hash rental already existed. NiceHash, Mining Rig Rentals, and a dozen sketchier cousins from 2013–2014 let you buy a stream of hash. Ryan knows that market from the inside; he wrote one of the original cloud hashing systems himself in 2013.

The difference, in his telling, is the traffic cop. Centralized rentals need a company to police buyers and sellers. Lumerin pushes that job into open-source nodes and smart contracts. The buyer’s Lumerin node watches delivery; the seller’s node watches payment. Under-deliver and the buyer’s node cancels. Underpay and the seller never routes hash. Tinker with the node to cheat and the economics break by design.

At the time of the interview the marketplace was live on Arbitrum mainnet, announced at the end of September around the same Prague trip where we first met. You download the Lumerin node from lumerin.io, fund Arbitrum ETH and LMR, and run an all-in-one interface that feels deliberately old-school Bitcoin: marketplace, wallet, proxy routing, and validation in one package.

Practical warning from Ryan: do not put this on a laptop that sleeps or updates overnight. You are streaming real hash. Turn the machine off and the contract dies.

His bigger pitch is almost aggressively cosmopolitan. North Korea, India, Nepal, Botswana: if you are a human on Earth you should be able to acquire Bitcoin. Governments may hate that framing. He does not.

Ethereum Rails, Bitcoin Soul & a Little Love and War

I could not resist needling him. Is this a secret plan to get Ethereum people mining Bitcoin? Because the UX is weird: hop onto an Ethereum L2, buy a token, deploy a smart contract, and only then rent Bitcoin hash rate.

Ryan leaned into it. Apart from tricks like this, there was no clean decentralized bridge from ETH or ERC-20s into native BTC. Buying hash and mining to your own address is one way to convert across ecosystems without a custodian, if you are patient.

I floated RGB or some future Bitcoin-native EVM layer so bitcoiners do not have to touch ETH to play. He agreed that is the endgame. Until then, hardcore bitcoiners look at Lumerin the way they look at many hybrid tools: they want to love it for attacking mining centralization, and hate it for living in web3. “A little love and war,” he said.

And hash power is only the first commodity. Any TCP/IP socket can ride the same pattern: AI compute, other PoW algorithms, file streams, telecom. The mining marketplace, in his view, is the demo.

Ryan is not much of a personal-brand guy. Follow Lumerin on Telegram and Twitter (lumerin.io), and eventually someone tags him. His name is Ryan Condron. If he were better at marketing himself, he joked, he might remember his own handles.

We cut the recording short because he had another call. Forty-two minutes was enough to go from garage extension cords to a global grid that treats hash like tradable electricity. Whether Lumerin becomes infrastructure or a footnote, the diagnosis still stings: Satoshi’s one-hash-one-vote world did not survive the toaster. The interesting question is whether control of those toasters can still be sold in pieces.