More on this book
Community
Kindle Notes & Highlights
Bitcoin and blockchains are not a technology story, but a psychology story: bubble economy thinking and the art of the steal.
You can make money from Bitcoin! But it is vastly more likely that you will be the one that others make their money from.
If you send bitcoins to a nonexistent address, they’re lost forever. If you send bitcoins to the wrong address, you can’t reverse it. Bitcoin security can be very technical, difficult and unforgiving;
This electricity is literally wasted for the sake of decentralisation; the power cost to confirm the transactions and add them to the blockchain is around $10-20 per transaction. That’s not imaginary money – those are actual dollars, or these days mostly Chinese yuan, coming from people buying the new coins and going to pay for the electricity. An ordinary centralised database could calculate an equally tamper-evident block of transactions on a 2007 smartphone running off USB power. Even if Bitcoin could replace conventional currencies, it would be an ecological disaster.
“Cyberlibertarianism” is the academic term for the early Internet strain of this ideology. Technological expertise is presumed to trump all other forms of expertise, e.g., economics or finance, let alone softer sciences.
The explicit promise of Bitcoin is that you can get in early and get rich – without even building an enterprise that’s useful to someone.
Pure commodities – gold and silver – haven’t done the job of money well for a few hundred years, and Bitcoin wants to be money but was set up to work like a commodity. Nakamoto put a strict limit on the supply of bitcoins: there will only ever be 21 million BTC.
Bitcoin ideology assumes that inflation is a purely monetary phenomenon that can only be caused by printing more money, and that Bitcoin is immune due to its strictly limited supply. This was demonstrated trivially false when the price of a bitcoin dropped from $1000 in late 2013 to $200 in early 2015 – 400% inflation – while supply only went up 10%.
Austrian economics, first put together in its present form by Ludwig von Mises (hence “Austrian”). Its key technique is praxeology, in which economic predictions are made entirely by extrapolating from fundamental axioms. It explicitly repudiates any sort of empirical testing of predictions, and holds that you can’t predict future behaviour from past behaviour even in principle, so testing your claims is meaningless:22
In practice, mining naturally recentralises due to economies of scale, so a few large mining pools now control transaction processing – and even though the cryptography is mathematically robust, the rest of the system is approximate, with attacks being a matter of how much economic power you can bring to bear.
Blocks in the blockchain were limited to 1 megabyte early on. But the blocks are now full – Bitcoin has reached capacity. This means a transaction may fail or be delayed for hours or days (if it isn’t just dropped), unless the user correctly guesses a large enough fee to get their transaction into the block. The Bitcoin community is unable to agree on how to fix this.
When mining rig manufacturer Butterfly Labs failed to deliver rigs on time, credit card and PayPal purchasers could do (and did) chargebacks; those who bought using bitcoins were out of luck.
There is no way on earth that Bitcoin could possibly scale to being a general utility. At 1 megabyte per block, the blockchain can only do a maximum of 7 transactions per second, worldwide total. Typical throughput in early 2017 was 2 to 4 TPS. Compare with the systems Bitcoin claims it can replace: PayPal, which ran about 115 TPS by late 2014;32 Visa, whose 2015 capacity was 56,000 TPS;33 even Western Union alone averaged 29 TPS in 2013.34
Bitcoin offered “equality” in that anyone could mine it. But in practice, Bitcoin was substantially mined early on – early adopters have most of the coins. The design was such that early users would get vastly better rewards than later users for the same effort. Cashing in these early coins involves pumping up the price and then selling to later adopters, particularly during the bubbles. Thus, Bitcoin was not a Ponzi or pyramid scheme, but a pump-and-dump. Anyone who bought in after the earliest days is functionally the sucker in the relationship.
Asset bubbles follow a standard progression: Stealth phase: The price of an asset is going up. Awareness phase: Some investors become confident, enthused by the rise. Mania phase: Popular buzz; media coverage. The public see these first investors and buy because others are buying, with the implicit assumption that there will always be Greater Fools to sell it on to. This is what makes a bubble: investing to sell to other investors. Someone will say that the old rules don’t apply any more. Blowoff phase: The old rules turn out to still apply. The bubble runs out of Greater Fools; prices
...more
“Stages in a bubble” by Jean-Paul Rodrigue, 2008.52 Bitcoin prices, January 2012 to January 2015. Totally no resemblance to the above. Data: coindesk.com
As a financial instrument born without regulation, Bitcoin quickly turned into an iterative exploration of precisely why each financial regulation exists. A “trustless” system attracts the sort of people who just can’t be trusted.
As of March 2015, a full third of all Bitcoin exchanges up to then had been hacked, and nearly half had closed.80 Since the exchanges are largely uninsured, unregulated and not required to keep reserves, depositors’ money goes up in smoke.
McCaleb saw the Slashdot post, tried and failed to buy some bitcoins, and thought an exchange would be a good idea. (Early Bitcoin core developer Martti Malmi had an exchange site, but it wasn’t very usable.96) He had run the “Magic: The Gathering Online Exchange,” a trading site for an online card game, for a few months in 2007, using the domain name mtgox.com;97 he quickly wrote some exchange software in PHP and reused the name because his girlfriend liked it.
They also had evidence from the Bitcoin blockchain – which, of course, contained a tamper-proofed record of every transaction ever conducted on it and which addresses were involved.117 Which is why Bitcoin is otherwise known as “prosecution futures”.
By the end of 2016, 75% of the Bitcoin hashrate was being generated in one building, using 140 megawatts133 – or over half the estimated power used by all of Google’s data centres worldwide at the time.
The Bitcoin block size is 1 megabyte per 10 minutes, which allows a theoretical maximum of 7 transactions per second.
users are in a blind auction, where they have to guess bigger and bigger fees in the hope of getting their transaction through. Transactions are routinely delayed hours or days, so many just get lost. (Only 57% of transactions are confirmed in the first hour; 20% never get confirmed at all, and are eventually dropped.191
When merchants do adopt Bitcoin, it tends not to result in a flood of business. Australian phone app MyBus, for local bus travel in Canberra, added Bitcoin as an option in March 2014, and had twenty-three transactions total by the time they removed it in January 2015. When they temporarily switched off the option for maintenance in September 2014, they received “about 30 emails from people asking for it to be reinstalled, which is odd because that’s more people than have actually used the feature.”220
Approximately 95% of on-chain transactions are day traders on Chinese exchanges;239 Western Bitcoin advocates are functionally a sideshow, apart from the actual coders who work on the Bitcoin core software.
But as the supplier of gambling trading facilities not available elsewhere, Bitfinex felt there was sufficient demand for their services that a drastic action would be considered acceptable to their users: rather than have some customers take a 100% loss, they assessed a 36% “haircut” on all customer deposits – including non-Bitcoin deposits. Depositors whose coins had been hacked would be compensated with money from depositors who hadn’t: “we are leaning towards a socialized loss scenario among bitcoin balances and active loans to BTCUSD positions.”258 The company would then try to trade its
...more
You might think that compensating your customers using money from other customers, while the owners don’t take a hit, would be against the rules in any reasonable financial system. Particularly as bankruptcies usually pay depositors and creditors first and equity holders last. But welcome to Bitcoin.
SingularDTV is a bizarre plan to fund a TV show about the Singularity in which a Caribbean island adopts Ethereum as its currency and Austrian economics works (this one gets its own section later in the book);
Iconomi is an index fund of other ICOs.
One Chinese “ICO” broke new barriers in market efficiency: you didn’t even need to put your ether into it yourself! Because the “white paper” contained malware that found your Ethereum wallet and emptied it. Now that’s a smart contract.323
In software testing, an oracle is any mechanism that determines if a test has passed or failed. The oracle problem is how to do this without costly human intervention.
Smart contracts make no sense as software engineering. You need a perfect bug-free program – but humans are really bad at coding without error. Programming to this extreme quality level is done by organisations like NASA for spacecraft, and it’s hideously slow and expensive.
This illustrated the final major problem with smart contracts: CODE IS LAW until the whales are in danger of losing money.
You can replace the term “distributed ledgers” with “shared Excel sheets” in about 90 percent of talk about blockchain and finance. – Tracy Alloway
The key problem with blockchain proposals for business are: Decentralisation is very expensive and doesn’t get you much, at the loss of efficiency and control. Recentralising immediately makes the system much more efficient. Your problem is pretty much always sorting out your data and formats, and blockchains won’t clean up your data for you.
If you have programmers, they probably save their code in Git, which is the closest I can think of to a useful blockchain-like technology: it saves individual code edits as transactions in Merkle trees with tamper-evident hashes, and developers routinely copy entire Git repositories around, identifying them by hash. It’s a distributed ledger, but for computer programs rather than money. What it doesn’t have is the blockchain consensus mechanism – you take or leave the version of the repository you’re offered. (I have had one “distributed ledger technology” developer admit his product was
...more
Git was released in 2005 and was based on work going back to the late 1990s; Merkle trees were invented in 1979. The good bits of blockchain are not original, and the original bits of blockchain turn out not to be much good.
Blockchain promises that it will let people who don’t trust each other work together. The trouble is that it does this only approximately, with startling inefficiency, and in a way that naturally recentralises to one or a few winners, as happened with Bitcoin.
The usual proposal to avoid this is to just start with central authority, at which point you probably shouldn’t be using a blockchain.
Exposing all your business data and back-office machinery to the whole Internet is obviously silly. So the next move was permissioned blockchains, for approved users only.
Their totally boss sci-fi TV series Singularity, no less! A worldwide economic collapse, as predicted by Austrian economics, leads to a fictional Caribbean island becoming the richest place in the world because it was first to adopt Ethereum as its currency. Then an artificial intelligence takes over the world, rendering the preceding plot meaningless. To be produced and distributed worldwide through the S-DTV portal!
Memoirs of Extraordinary Popular Delusions and the Madness of Crowds by Charles Mackay was first published in 1841 and remains the best book available on economic bubble thinking.434
