I do agree that it is really hard to take Rubini seriously.
Much of that has to do with the way he expresses itself: he's really obnoxious to listen to and one has to make a great effort to get past that 100-foot wall of aggression to try and understand if there's any value in his arguments.
I have honestly tried, and once you peel off the tone and posturing, I must confess there isn't much to be found.
Not to forget that top epidemiologists demand to stop the devastating measures immediately because our data shows that they cause much more harm than good.
As far as I understand there is no such thing like ASIC resistance. Whenever there is an incentive people will build ASICs. It doesn’t matter how complicated your algorithm is.
Right, but this approach solves that by not defining any hashing algorithm at all. Every block includes its own hash algorithm source code. Every node decides for itself whether to accept the hash algorithm and how much to value it. If people suspect an algorithm is getting ASIC'd, they can devalue it (if they want -- maybe the majority of nodes want ASICs for some reason, and that's fine too). The point is to let the users to choose the algorithm, not a core group of developers.
1) If every block includes source code for an arbitrary hashing algorithm, do you expect arbitrary nodes to run arbitrary code to synchronise with your block? How do you know if some block provider's hashing algorithm terminates? How does your blockchain become trustless if hashing involves arbitrary code execution?
2) An essential property of hashing algorithm is uniform distribution. Future block hashes depend on past block hashes. In PoW this implies compute difficulty scales dynamically. If a block provider gets to choose their hashing algorithm arbitrarily, they could pick a non-uniform hashing algorithm that is easy for themselves and hard for others. In PoS this often affects who gets picked as the next block provider; having control over the hashing function in either case means you can pick yourself again.
If your claim is to let market forces pick the most trusted hashing algorithms, why not just settle on a fixed-but-extensible list of hashing algorithms that are known to work well today?
Making governance a part of the protocol is all the hype.
But there has to exist a protocol with boundaries.
1) A node _could_ run arbitrary code, but unless it trusts the code's provenance then it should use a white-list as you mention.
> why not just settle on a fixed-but-extensible list of hashing algorithms that are known to work well today?"
Sure, that's what a whitelist will be, but with one major difference: each node gets to pick its own whitelist, there's no built-in centralization via the "one true source code."
2) Yes, miners will choose hashing algorithms that are easy for themselves and hard for others, but the nodes get to value each block from their own perspective. If they think a miner is using a weird hash function, they can penalize it or just blacklist the function.
Check out the RandomX PoW algorithm for Monero. To get around the ASIC resistance issue, they instead targeted a particular component (the CPU) and designed the algorithm around it.
It works really well, FPGAs and ASICs seem to have been stopped in their tracks.
The WHO recommends to label everyone as a c19 death who was tested positively within the last month. They say they explicitly want to overestimate the number of deaths because they don’t want to underestimate it to be on the save side.
Additionally, the relatively high false-positive rate leads to an even higher overestimate.
Also aren’t there massive financial incentives for clinics to label someone as a c19 death? IIRC an ICU bed for a c19 patient costs 5-10x more per day than for other patients.
This article is a preprint and has not been peer-reviewed [1]. It reports new medical research that has yet to be evaluated and so should not be used to guide clinical practice.
[1]Readers should therefore be aware that articles on medRxiv have not been finalized by authors, might contain errors, and report information that has not yet been accepted or endorsed in any way by the scientific or medical community.
Ioannidis is one of the most cited researchers ever. In particular he’s famous for debunking bogus science with accurate statistics. He’s absolutely among the top epidemiologists in the world.
Bro, it’s ridiculous to try to downplay his analysis without even trying to understand it. XD
I see a number of problems with the methodology (eg. excluding data points after an arbitrary one-week-after-study-midpoint which author acknowledges as arbitrary, doing an averaging/smoothing of results per location first, then doing a global calculation, which compounds the error).
But the numbers do seem reasonable. Still, the difference with flu is that there is no protective vaccine to give to people at higher risk, nor do we know exactly what constitutes the higher risk, nor do we understand exactly how transmissions happen or why they do not happen.
Also note that we are already in a largely manipulated situation to be able to judge any "organic" data (we've been social distancing, hugely reduced international travel, and a stronger dedication to doing business online). While some will point at the number of people still behaving normally, I always point at the number of people who are not, reminding that this is a "game of statistics".
Finally, to me personally, global IFR and R0 numbers do not matter. I do not care how many people are affected, but how much risk does the infection pose to my close family and friends and me? It seems there is a strong correlation between heart-related diseases and coronavirus deaths, so I would like to see numbers for "extra coronavirus deaths over heart-related deaths" for an "unmanaged" first month in the US (eg. April had ~60k people, May-July another 90k). My parents, both with heart problems, barely made it through flu season a number of years back (they've been on flu vaccines since). Heart-related deaths in US average out at 600k/year or 50k/month (did not find per month averages to account for seasonal spikes). To be honest, I expect to see a huge decrease in heart-related deaths in these months, but I couldn't find anyone doing a study along those lines (or well, simply having access to current numbers)! Still, none of that helps me until there is (safe! still a year away likely) medicine to prevent or treat COVID19.
