Blockchain is the world's worst database, created entirely to maintain the reputations of venture capital firms who injected hundreds of millions of dollars into a technology whose core defining insight was "You can improve on a Ponzi scam by making it self-organizing and distributed; that gets vastly more distribution, reduces the single point of failure, and makes it censorship-resistant."
That's more robust than I usually phrase things on HN, but you did ask. In slightly more detail:
Databases are wonderful things. We have a number which are actually employed in production, at a variety of institutions. They run the world. Meaningful applications run on top of Postgres, MySQL, Oracle, etc etc.
No meaningful applications run on top of "blockchain", because it is a marketing term. You cannot install blockchain just like you cannot install database. (Database sounds much cooler without the definitive article, too.) If you pick a particular instantiation of a blockchain-style database, it is a horrible, horrible database.
Can I pick on Bitcoin? Let me pick on Bitcoin. Bitcoin is claimed to be a global financial network and ready for production right now. Bitcoin cannot sustain 5 transactions per second, worldwide.
You might be sensibly interested in Bitcoin governance if, for some reason, you wanted to use Bitcoin. Bitcoin is a software artifact; it matters to users who makes changes to it and by what process. (Bitcoin is a software artifact, not a protocol, even though the Bitcoin community will tell you differently. There is a single C++ codebase which matters. It is essentially impossible to interoperate with Bitcoin without bugs-and-all replicating that codebase.) Bitcoin governance is captured by approximately ~5 people. This is a robust claim and requires extraordinary evidence.
Ordinary evidence would be pointing you, in a handwavy fashion, about the depth of acrimony with regards to raising the block size, which would let Bitcoin scale to the commanding heights of 10 or, nay, 100 transactions per second worldwide.
Extraordinary evidence might be pointing you to the time where the entire Bitcoin network was de-facto shut down based on the consensus of N people in an IRC channel. c.f. https://news.ycombinator.com/item?id=9320989 This was back in 2013. Long story short: a software update went awry so they rolled back global state by a few hours by getting the right two people to agree to it on a Skype call.
But let's get back to discussing that sole technical artifact. Bitcoin has a higher cost-to-value ratio than almost any technology conceivable; the cost to date is the market capitalization of Bitcoin. Because Bitcoin enters through a seigniorage mechanism, every Bitcoin existing was minted as compensation for "security the integrity of the blockchain" (by doing computationally expensive makework).
This cost is high. Today, routine maintenance of the Bitcoin network will cost the network approximately $1.5 million. That's on the order of $3 per write on a maximum committed capacity basis. It will cost another $1.5 million tomorrow, exchange rate depending.
(Bitcoin has successfully shifted much of the cost of operating its database to speculators rather than people who actually use Bitcoin for transaction processing. That game of musical chairs has gone on for a while.)
Bitcoin has some properties which one does not associate with many databases. One is that write acknowledgments average 5 minutes. Another is that they can stop, non-deterministically, for more than an hour at a time, worldwide, for all users simultaneously. This behavior is by design.
I can go on, and probably will some other day. This is a bit of a hobby for me.
Blockchain is the world's worst database, created entirely to maintain the reputations of venture capital firms who injected hundreds of millions of dollars into a technology whose core defining insight was "You can improve on a Ponzi scam by making it self-organizing and distributed; that gets vastly more distribution, reduces the single point of failure, and makes it censorship-resistant."
That's more robust than I usually phrase things on HN, but you did ask. In slightly more detail:
Databases are wonderful things. We have a number which are actually employed in production, at a variety of institutions. They run the world. Meaningful applications run on top of Postgres, MySQL, Oracle, etc etc.
No meaningful applications run on top of "blockchain", because it is a marketing term. You cannot install blockchain just like you cannot install database. (Database sounds much cooler without the definitive article, too.) If you pick a particular instantiation of a blockchain-style database, it is a horrible, horrible database.
Can I pick on Bitcoin? Let me pick on Bitcoin. Bitcoin is claimed to be a global financial network and ready for production right now. Bitcoin cannot sustain 5 transactions per second, worldwide.
You might be sensibly interested in Bitcoin governance if, for some reason, you wanted to use Bitcoin. Bitcoin is a software artifact; it matters to users who makes changes to it and by what process. (Bitcoin is a software artifact, not a protocol, even though the Bitcoin community will tell you differently. There is a single C++ codebase which matters. It is essentially impossible to interoperate with Bitcoin without bugs-and-all replicating that codebase.) Bitcoin governance is captured by approximately ~5 people. This is a robust claim and requires extraordinary evidence.
Ordinary evidence would be pointing you, in a handwavy fashion, about the depth of acrimony with regards to raising the block size, which would let Bitcoin scale to the commanding heights of 10 or, nay, 100 transactions per second worldwide.
Extraordinary evidence might be pointing you to the time where the entire Bitcoin network was de-facto shut down based on the consensus of N people in an IRC channel. c.f. https://news.ycombinator.com/item?id=9320989 This was back in 2013. Long story short: a software update went awry so they rolled back global state by a few hours by getting the right two people to agree to it on a Skype call.
But let's get back to discussing that sole technical artifact. Bitcoin has a higher cost-to-value ratio than almost any technology conceivable; the cost to date is the market capitalization of Bitcoin. Because Bitcoin enters through a seigniorage mechanism, every Bitcoin existing was minted as compensation for "security the integrity of the blockchain" (by doing computationally expensive makework).
This cost is high. Today, routine maintenance of the Bitcoin network will cost the network approximately $1.5 million. That's on the order of $3 per write on a maximum committed capacity basis. It will cost another $1.5 million tomorrow, exchange rate depending.
(Bitcoin has successfully shifted much of the cost of operating its database to speculators rather than people who actually use Bitcoin for transaction processing. That game of musical chairs has gone on for a while.)
Bitcoin has some properties which one does not associate with many databases. One is that write acknowledgments average 5 minutes. Another is that they can stop, non-deterministically, for more than an hour at a time, worldwide, for all users simultaneously. This behavior is by design.
I can go on, and probably will some other day. This is a bit of a hobby for me.