California Gathering of Bitcoin Miners and Devs (07/30/2016). Highlights by Tuur Demeester
Thanks to @tuurdemeester for going through the transcript and highlighting the important parts.
2/ Concerns about #ethereum: economy rules keep changing, cause contract malfunction, blockchain growth exponential.
Cost of spinning a Bitcoin node more than blockchain download time. “Maybe in 5 yrs, nobody would be able to download the blockchain.”
We’re building “marine shipping containers for bitcoin mining”, “You just need the power at your location” … “immersion cooling”
Social attack vectors: “Everything from last few years has helped our immune system. We’ll be facing larger adversaries in the future.”
Lightning network as decentralized exchange: “I could send a lightning payment on bitcoin that releases a lightning payment on litecoin”
“fungibility is a big problem for Bitcoin”… “if we had a space-efficient secure protocol to do fungibility, we would do that right now”
Bitcoin scaling helps with fungibility: “if we don’t broadcast as much information … you broadcast less data”
Lightning: “not just for bitcoin, …. any coins that exist can be connected together into lightning network … might help fungibility”
Forking: unlike Linux, “bitcoin consensus protocol needs only 1 chain otherwise we have problems like ETH is presently experiencing”
old approach “if I can control all the mining hardware in my own hand, then I can benefit most from this control” is failing.
new approach: “only keep a small percentage of the machines for ourselves” (this comment likely from Bitmain)
“I do work on lightning network. My intent is to help make bitcoin great again.”
Antpool’s hashrate is about 14% of the network. “76% of our customers come from outside of China. United States takes half of it.”
74.8% of Bitmain’s machine orders come from Chinese nationals. “the Chinese customers are more likely to build larger farms”
“American customers buy only one or two units of machine … Lots of our Chinese customers are large farms like 1 or 2 megawatts.”
Mining in China vs US: Build time: 1 month vs. 3 CapEx/MW $50k vs. $300k Cheapest power $0.04/kwh vs. $0.02/kwh
Electricity in China: “less hydro than US” … “oligopoly” … “against the rules to make deal with a power company directly”
ASIC development for mining: “Bitcoin experienced technology improvement journey of 13 year semiconductor technology in about 3 years.”
Bitcoin mining: 20x technology improvement in 3 years time
Commoditization of Bitcoin mining: “Yes we’re hitting a wall” “improvement of mining tech will have to wait on improvement of chip tech”
“potential that Bitcoin mining will shift [from China] to the West because regulation, the legal system and the power cost”
#Bitmain: “Mining pools centralizing factor, but [by] open-sourcing, It’ll become easier for others to build a mining pool.”
“It’s very difficult right now to setup the existing mining pool software.” “P2Pool is not good … never was”
There is tech that could be developed here where users can pool their income but keep transaction selection decentralized. Multi-PPS?
“The mining pools joining together to share their income — increases network performance … Combining pools is the next logical step.”
“In Ukraine, our partners have already designed a Bitcoin miner that works like a heater for a house. It creates warm water.”
Trade off: “efficiency of 1M chips, not in single miner but distributed across the world, will be lower bc of delay in communication.”
Effects of halving: “80% of 28nm chip machines will be shutting down” (Cutting edge ASICs from April 2014, now already obsolete)
Miners as heating devices: “With 10 nm ASICS: difficulty increase more linear, recoup investment in 2–3 months + benefits after that.”
32/ This next part is a new session, titled “Future outlook of Bitcoin”.
Paypalisation risk: “One potential outcome for bitcoin is for exchanges to become banks or bought by banks or incorporated into banks.”
Positive outcome for Bitcoin in 5–10 years: “the same amount of distribution and value as gold.”
Maintaining Bitcoin’s unique value proposition
“A wild Bobmon has appeared in the room. It’s very important to capture this pokemon.”
37/ Next section in the transcript is called “Lessons from the #Ethereum hard-fork”
#TheDAO fork: “I was surprised that they did not have an improvement proposal for it. They just did a hard fork without writing it up.”
On #Ethereum’s scripting language, Solidity: “should be improved” (May be a transcript misinterpretation on my part, open to feedback)
“If you change something in Solidity … [existing smart contracts] might stop working … no continuous support of the same system.”
“Changing from proof of work to proof of stake changes the economics of the system, all the rules change and it will impact everything.”
ETH client: “[to get latest state] you should [process] all the smart contracts. [They use] solidity, consume a lot of CPU power”.
If blockchain keeps growing “you would not be able to download the blockchain and synchronize new nodes”, and if during that time …
… “smart contracts would grow bigger, then regular node wouldn't be able to process all smart contracts” — causing need to Hard Fork.
Each scaling bottleneck is new pandora’s box: Bad HF = death of ethereum Good HF = changing rules of whole system, burdening the user
#Ethereum doesn’t think about the continual loss of trust caused by testing the network as they go and modifying the rules.
#TheDAO: “response time seemed slow to me. … Past Bitcoin network incidents, response time was much faster I think.”
1.5 yrs ago, replay attack mitigation was discussed in Bitcoin-dev chat channel w/regards to bitcoin, bc of Mike & Gavin’s HF proposals.
Bitcoin’s UTXO model (as opposed to ETH’s account model) can more easily be protected against post HF replay attacks.
ETH devs had a pre HF convo with coinbase about replay, position was that there would’t be two surviving chains so don’t worry about it.
Ethereum’s failure mode: overconfidence. Vs Bitcoin: bit of over-engineering, good because one can’t predict each failure mode.
Question: Post Hard Fork, should the minority chain be allowed to exist? (it’s possible to program in self erasure of least work chain)
“large economic incentive to encourage splitting of assets. Poloniex made $200k in fees in first several hours after listing ETC.”
Exchange’s ETC loss: “Ethereum team might have liability … bc they are selling a piece like a commercial company” … also “premine”.
Bitcoin HF: unsatisfied chain “might start taking legal action” … “[Bitcoin core devs & miners] would all be liable to some extent.”
Easy to make a “forced fork” where minority chain cannot exist. Yet legal action could ensue: “you’re preventing us from transacting”.
ETH community says split is safety mechanism: “If majority of miners want to do something bad, network can continue on without them.”
Minority fork prevention in Bitcoin: “Even if we could make it impossible, we would lose an argument for Bitcoin’s long-term safety.”
“good to establish the precedence that exchanges in the event of a fork, have an obligation to give coins on the old fork back”
Soft forks “impose cost on all the users of bitcoin” Imposing too much too quickly => users’ll ignore core, adopt strange altcoins, etc.
BTC community’s lesson should be to adopt standard duty of care: “are devs moving too fast? Are there good reasons for this change?”
“ Has ethereum scared users away from cryptocurrency in general? I think it has some, at least.”
Core dev: “If I say that Ethereum has bad design … ETH people would throw more rocks. Enough people are throwing rocks at me already.”
“Lot of schadenfreude … Ethereum’s damage hurts us all to a very significant level. At the same time I could have warned people more.”
“If cryptocurrency as a whole doesn’t mature … the result will be either failure, or regulatory pressure with enormous cost for all.”
“Can’t function in an environment where developers as a whole can’t operate in the public. … pseudonymous … nothing would get done”
“Notice how outside the room it does not say ‘bitcoin meeting’. FBI knows. I bet your plane ticket says.”
“Ethereum debacle is bad for Bitcoin bc it shows that some cryptocurrencies are not trustable. Public doesn’t know the difference.”
70/ Bitcoin product differentiation vs. Ethereum: Not “we don’t hard fork ever”, but rather “Bitcoin has vigorous duty of care”.
“Once we we take out the human element, then bitcoin is really immutable.” — AI’s writing code — Automated mining facilities
We’re taking the human element out of Bitcoin: script validation is law. Lesson from ETH: saying “code is law” doesn’t make it true”.
“Fungibility boosts immutability because if you can’t target the coins to take back, then you can’t even make a proposal to undo it.”
“if you build a system where it’s near impossible to reverse [transactions], people are more diligent abt what they do with their money”
“ You want … sufficient diligence such that there is no demand for reversal in the first place. … could be done by education.”
HF to increase block size: hard to know how many people want it. In ETH, only 20% or so voted. Also: people have now seen HF go wrong.
Inadvertent burn: “The outcome for Ethereum is no where near as bad as it would be for Bitcoin, which is actually used by people.”
ETH Foundation releasing HF gave it official air, gave it more traction than it would have had organically. Pools ignored user votes.
People are often heavily influenced by who they’re told will win. In Ethereum, Foundation releasing HF => The Buterin effect.
“a lot of miners had no opinion either way, mined [ETC] because they could make profit. Something to consider for Bitcoin’s future”
Soft fork upgrades like SegWit are non-repeatable. HF block size increase: indefinite, could attempt to “create a new coin every day”.
Ethereum does live network testing. Indeed some things that you can only learn in the production network. Though testnets do have value.
The fact that nobody would run testnet with lots of 8 MB blocks, doesn’t mean that nobody would run Bitcoin with 8 MB blocks.
Ethereum devs are all getting paid in Bitcoin. (USD denominated salaries)
Exchanges Post HF: Say you’re a gold depository. And you’re switching to lead. (…) at minimum you should deliver the gold bars.
Exchanges Post HF: Better example are stock splits. It’s clear that you should give both parts back, after a stock split. It’s property.
“One thing Ethereum did well was to make lots of positive PR, continued to taunt success, while things were failing in the background”
For Bitcoin, we need balance between frank technical talk and positivity, fex companies creating joint positive marketing.
“Before #Ethereum HF, many people had illusion: maybe one is the genuine ETH — But after we just see there’s actually no genuine chain.”
“If Bitcoin HF’d to unlimited size, I’d quit — Block size determines decentralization & if people can participate running the nodes.”
“If Bitcoin’s fungibility breaks, people would … move on. Block size can break fungibility, and that’s why it’s controversial.”
“There’s a common argument that with a soft-fork there is less opportunity for people to disagree.” Yet miners can disagree just fine.
Not running the SF de facto means blocking transactions, for ex SegWit tx. Question is how would users respond to this miner behavior?
“Ethereum sets HF precedent — Only two possibilities: either you accept there will be multiple chains, or accept that you can attack.”
“51% attacks to suppress an altcoin won’t be successful; it’ll change its rules to suppress the attacks using security defense attacks.”
Tuur Demeester will be posting part 2 through his medium