Read on Twitter
Hi, I'm Mad! Welcome to my blockchain rant. Specifically: for genomes, health, and personal data.
"Why aren't you using blockchain?" they ask me. I'll tell you why!
Maybe you can RT this – or quote specific points – to inoculate other folks against fuzzy-brained BS
1/16
"Why aren't you using blockchain?" they ask me. I'll tell you why!
Maybe you can RT this – or quote specific points – to inoculate other folks against fuzzy-brained BS
1/16
1. You aren’t storing the data in the blockchain.
The blockchain is a distributed ledger. Everyone has a copy of the ledger. It’s insanely expensive to store data in it. 2/16
The blockchain is a distributed ledger. Everyone has a copy of the ledger. It’s insanely expensive to store data in it. 2/16
(So the blockchain will only record very small things. Tiny fragments of data that could be used to identify or cryptographically verify the integrity of a file.) 3/16
2. The blockchain doesn’t secure your data.
It’s not encrypting anything. It’s just storing tiny pieces of data, permanently. That’s what makes it special. Once something is stored, nobody can change it. 4/16
It’s not encrypting anything. It’s just storing tiny pieces of data, permanently. That’s what makes it special. Once something is stored, nobody can change it. 4/16
3. You can store data verification – but why?
Data providers can provide hashes for data verification already. The only reason to store it on the blockchain is if you trust no one – if you need a rock solid verification for something, that can never be changed. 5/16
Data providers can provide hashes for data verification already. The only reason to store it on the blockchain is if you trust no one – if you need a rock solid verification for something, that can never be changed. 5/16
4. Permanent data verification is useful for specific use-cases – like "currencies".
If there is value to be claimed, we worry about forgery. 6/16
If there is value to be claimed, we worry about forgery. 6/16
5. So "blockchain for X" is almost always about "monetizing X".
It’s not about data security, it’s not about making the data better quality or universally useful. It’s about creating a marketplace for the data. 7/16
It’s not about data security, it’s not about making the data better quality or universally useful. It’s about creating a marketplace for the data. 7/16
6. If it’s not about money, it’s probably overkill.
Ensuring data integrity? Reporting updates? Blockchain isn’t needed to do this! If it’s not about monetization, you can usually ask "How else could we solve this?" and get a better answer. 8/16
Ensuring data integrity? Reporting updates? Blockchain isn’t needed to do this! If it’s not about monetization, you can usually ask "How else could we solve this?" and get a better answer. 8/16
(If you just need a database, use a database! For pity's sake, don't use blockchain to reinvent databases.) 9/16
7. A blockchain isn’t needed to monetize things!
Centralized brokers sell stuff all the time. To put a more specific point on it: blockchains enable decentralized monetization. 10/16
Centralized brokers sell stuff all the time. To put a more specific point on it: blockchains enable decentralized monetization. 10/16
8. Decentralizing your personal data sounds pretty dangerous (part 1)
What if you lose your own encryption keys? In the decentralized world implied by blockchain, there is no "Forgot password?" link. 11/16
What if you lose your own encryption keys? In the decentralized world implied by blockchain, there is no "Forgot password?" link. 11/16
8. Decentralizing your personal data sounds pretty dangerous (part 2)
Centralizing data can also help with security: you have the power to delete data, and potentially correct an otherwise catastrophic leak. 12/16
Centralizing data can also help with security: you have the power to delete data, and potentially correct an otherwise catastrophic leak. 12/16
9. Monetizing data is weird (part 1)
The value of data is contextual. Monetizing data is, in itself, a tricky proposition. Its value depends on the situation or presence of other data – certain types (e.g. health survey) or in sufficient quantities (e.g. >10,000 genomes). 13/16
The value of data is contextual. Monetizing data is, in itself, a tricky proposition. Its value depends on the situation or presence of other data – certain types (e.g. health survey) or in sufficient quantities (e.g. >10,000 genomes). 13/16
9. Monetizing data is weird (part 2)
Data isn’t a finite good. Data can also be copied indefinitely. Monetization presumably requires exclusivity, which must be artificially imposed. (Think about the market for MP3s.) 14/16
Data isn’t a finite good. Data can also be copied indefinitely. Monetization presumably requires exclusivity, which must be artificially imposed. (Think about the market for MP3s.) 14/16
10. Just because it’s precious to you doesn’t make it precious to others.
For others, your data is interchangeable – it can be worth a lot less than you imagine! Individuals may be unmotivated to "sell" at realistic values, and a user-centered marketplace fails to thrive. 15/16
For others, your data is interchangeable – it can be worth a lot less than you imagine! Individuals may be unmotivated to "sell" at realistic values, and a user-centered marketplace fails to thrive. 15/16
11. It sounds like hype.
Blockchain for genomes, health, and personal data sounds like hype to me – and it probably sounds like hype to you! But now, I hope I’ve given you reasons that describe why your skepticism may be warranted.
16/16
Blockchain for genomes, health, and personal data sounds like hype to me – and it probably sounds like hype to you! But now, I hope I’ve given you reasons that describe why your skepticism may be warranted.
16/16
12. An addendum regarding "smart contracts"…
Smart contracts are basically fully automated escrow, and I can't recall anybody saying escrow is in dire need of full automation.
17/16
Smart contracts are basically fully automated escrow, and I can't recall anybody saying escrow is in dire need of full automation.
17/16
