@GidMK I've been ranting against cryptocurrency for quite a while, and I fully agree with every word he says.
@GidMK TL/DR: blockchain is a database that can store any data. Nothing more. But it's distributed between many parties, who you can't trust. So unlike a normal database, you have to go to extreme lengths to try to prove that they're not lying to / cheating on you.
@GidMK With cryptocurrencies, they're using it like a database of VISA transactions of digital coins, and also generating a small number with each transaction processed. You could of course do this with any database, but distributed you must assume everyone might be lying.
@GidMK Unlike a centralized database, where the central authority can fix problems as needed, everything written to the blockchain is permanent. If it's money, there's no reverting a transaction. If it's software, there's no fixing a bug. Etc.
@GidMK Cryptocurrency advocates argue that this is an advantage. Permanence is good. Central authorities are evil. But rarely are real-world databases designed to be immutable, because that's rarely a characteristic that's *actually desirable*.
@GidMK Another aspect is that all data stored in a blockchain is publicly accessible. You can insert some second inner layer to hide it (most cryptocurrencies don't), but the default is "everyone can see everything", while central servers default to "everything is private".
@GidMK (But of course there are nice cheap laundering services called "mixers" that randomly shuffle up coins between different addresses so criminals can separate their expenditures from the sources of their illicit funds)
@GidMK It's possible to design cryptocurrencies which follow anti-laundering/anti-terrorism/etc and know-your-customer regulations. But the anti-authority attitude of most crypto-advocates are generally staunchly against this.
@GidMK (Of the relatively small portion of crypto advocates who actually make non-speculative financial transactions, it's amazing the percentage who will openly admit to you that they do so to avoid taxes, buy illegal things, etc)
@GidMK In order to avoid the central authority, again, you have to implement proofs of having committed a certain amount of things that cost money to run, which all add together each time a transaction is processed, proving that making the chain took great resources.
@GidMK Usually these resources are processor power and electricity (ASICs). Sometimes it's that plus memory (GPUs). Occasionally it's disk storage (but with a lot of CPU/GPU overhead). In all of them, you're proving you wasted a ton of power, hardware, and made e-waste.
@GidMK Central authorities don't have to do any of this, of course; they have "proof of authority". So their footprint is *many* orders of magnitude smaller per transaction.
@GidMK The theoretical exception for blockchain is "proof of stake", where coins holders are randomly selected as validators, with their coins at stake for validation that others disagree with. Security risks are less well quantified vs. Proof of Work.
@GidMK Even under proof of stake resource consumption per transaction is still vastly higher than a normal database with a central authority. Just not "numerous orders of magnitude more".
@GidMK TL/DR: Blockchain is a massively inefficient immutable publicly-visible database for people who hate central authorities with a passion, and would rather work with peers they can't trust.
@GidMK Governments, for their part, can make blockchain databases they control, too. And entities seeking to evade governments can also used central databases not under said governments' control. Central vs. distributed does not define the ability to evade censorship/control.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
A slight Sunday decline in UK cases to 37011 corresponds to a slight drop in the doubling time to 37,8d.
In the UK, the decision to suspend all COVID controls in the House of Parliament has been met with a chorus of condemnation from doctors and medical experts, including...
On the other side of the spectrum, outrage has been thrown at comments by vaccine minister Nadhim Zahawi that while...
parental consent will be normally required for children to get vaccinated if the government goes through with their plan to overrule the JCVI recommendations, objecting teens deemed "competent" to make choices will be able to overrule their parents.
With a mild Saturday decline to 37578, UK doubling times drop down to 39,5d, just over a month. Hospitalization and mortality trends continue their unending upslope.
In self-inflicted chaos news, the JCVI's recommendation against vaccinating 12-15 year olds - a decision that...
sets them apart from expert groups around the world - has now gotten a rebuke from a member of SAGE, the government's *other* pandemic advisory group. While JCVI portrayed *vaccination* as disrupting schools, Sage's Prof. Edmunds warns COVID will do that:
Johnson looks to plan to push ahead on vaccinations for 12-15yos regardless of the JCVI's statement, but will now face more parliamentary opposition. In a twist, JCVI's Prof. Harden is publicly saying that parents should be given an option to vaccinate:
Their opinion, which references only three non-peer-reviewed preprints, is full of misleading and outright wrong statements - for example:
Anyone should immediately see the problem of using the *whole population* as the denominator rather than the rate per infection (when talking about allowing all children to get infected), but the numbers don't even match their preprints, and are more to the point, impossible.
32367 cases on a Saturday brings the centre-averaged doubling time down to slightly to 18 days. UK hospitalizations continue the expected exponential trend.
It's been previously discussed that Delta's effectiveness is by and large not through any clever immune-evasion tricks,...
... but rather largely through simply being a more "fit" virus due a higher binding affinity for ACE2. New research continues to show how fit it is:
Over the course of the pandemic, the mean time between exposure and PCR-detectable levels of ...
... virus has fallen from 6 to 4 days. More concerningly however, the detectable levels of the virus were 1260 times (not percent... times) higher at the first detection than early in the pandemic. Delta, quite simply, surges to high viral loads extremely quickly.
35707 UK cases today drops the doubling time down a notch to 22,7 days. Despite 4 of 10 of England's ambulances already being overloaded and on "Black Alert" (something we only know thanks to leaks):
... government not insisted on lifting the brakes on 19 July, whether the continued progress of vaccination would be able to dull this wave before it gets too bad. A 3-week doubling time is a significant improvement, after all.
But I guess we'll never know.
With all but a handful of European countries back into case growth, Germany - one of the more slow-growing European countries - has declared all of Spain a COVID risk area and requires that travelers provide a negative test result to avoid quarantine.
Lots going on in the charts today, in this period of calm before the brakes get lifted on the 19th.
The good news first: meagre growth to 32551 cases on a Thurday boosts the doubling time to 26,1 days.
The bad news: the death trend continues to rise further above the...
... hospitalization trend, nearly keeping track with the case trend.
We finally got age breakdown data for June, and there is some good news: the growing incidence rate of childhood hospitalizations relative to cases has reversed and is back to winter levels. While it's not...
... clear why this has happened, it's certainly welcome.
Vaccination in June switched to primarily focusing on the 18-54yo cohort, and this can be clearly seen in their declining share of cases in graph 2. The 55+ cohort continued to decline as immunity built, but appears to...