mrb

58 ms·

631.
by mrb 7y ago
I agree, getting the flu shot is very important. My whole family gets it every season.
632.
by mrb 7y ago
I suppose his point is that the coronavirus overwhelms hospitals much, much more seriously than the flu does.
633.
by mrb 7y ago
Simple, check this log scale chart of COVID-19 cases excluding China: https://pbs.twimg.com/media/ETgWL8rU8AA3OML?format=png&name=... They multiply by 10 every 14 days. So assuming governments did not take any acti
634.
by mrb 7y ago
I hate credentialism like this. It's anti-intellectual. There are many smart people on HN who are perfectly capable of reading and understanding research literature in an area that is NOT their primary expertise, and who can make half-
635.
by mrb 7y ago
French sources specifically mention severe cases of covid-19 in "young and healthy" patients (eg. see https://lefigaro.fr/sciences/coronavirus-alerte-sur-l-ibupro... ) but this important piece of information
636.
by mrb 7y ago
" Being consistent with incorrect studies is nothing to brag about. ” So ALL(!) the studies referenced in WHO situation reports are "incorrect"? The WHO couldn't/didn't find any other studies that were "co
637.
by mrb 7y ago
Well, look at this chart which speaks volume about how successful China is in containing the outbreak, with new cases per day decreasing dramatically since February: https://pbs.twimg.com/media/ESywcEKUwAADhcQ?format=pn
638.
by mrb 7y ago
I hate to suggest that but the only short-term solution that seems to work is large-scale extreme response measures, just like China. Here is evidence why it works: new cases/day decreasing dramatically since February: https:/
639.
by mrb 7y ago
16 days ago I wrote: " The reason the WHO is not declaring the coronavirus a "pandemic" is likely to avoid causing a panic at this point. They will only use the P-word once it is clear to everyone that it is one. " http
640.
by mrb 7y ago
The growth rate in South Korea is low (1.05) because they are doing fewer tests, as they are finishing testing the sect members, as KCDC explained. IOW it does not reflect the true growth of the outbreak, but is an artifact of testing slo
641.
by mrb 7y ago
A lockdown has worked in only one place: China. And it worked because the measures were extreme: not allowing people to leave their homes (except once a week to go shopping), banning cars on the roads, closing workplaces, etc. No country is
642.
by mrb 7y ago
Just looked at the ATX12VO spec and it mentions the max current per pin, page 12: " the guideline to follow is 6-8 Amps per pin "
643.
by mrb 7y ago
Actually the limiting factor is the amperage rating of the molex terminals ("pins") that make the physical contact. I have expertise in this area back from the days of Bitcoin GPU mining where I built custom cables that I crimped
644.
by mrb 7y ago
Divergent warps are still a huge problem (but SLIDE doesn't have this problem AFAIK). Uncoalesced reads are not a problem severe enough to make GPUs underperform CPUs. Or, said another way, uncoalesced reads come with a roughly equally
645.
by mrb 7y ago
Reads shorter than 64 bytes on a CPU all cost you the same: a packet of 64 bytes on the memory bus, because that's the atom size of modern CPU's DDR4 memory controllers... On GPUs the atom size is 32/64 bytes. So GPUs are alw
646.
by mrb 7y ago
GPUs would still be faster than CPUs. You describe them as high-latency but their memory latency is comparable to CPUs. That's why ethash mining or equihash mining (workloads bottlenecked by short ≤32-byte random memory reads) is stil
647.
by mrb 7y ago
" Your example is a false comparison, generating a hash is very different from pointer chasing the address generated by the hash. " That's not true at all in the case of ethash, where the running time is dominated by waiting
648.
by mrb 7y ago
As someone who ported memory-bound workloads to GPU, I say SLIDE appears it would run even faster on GPU than on CPU. I skimmed the paper and SLIDE is described as a memory-bound workload, specifically random memory accesses to 2-10GB of da
649.
by mrb 7y ago
A black swan is a rare event. It's appropriate for COVID-19, as its death toll is either already higher than some of the outbreaks you listed, or is expected to surpass them in the near future. Bill Gates called COVID-19 a once-in-a-ce
650.
by mrb 7y ago
" First, COVID-19 does not transmit as efficiently as influenza, from the data we have so far. " Ok let's look at the data he summarizes in the next sentences: " With influenza, people who are infected but not yet sick a
651.
by mrb 7y ago
The coronavirus is a magnitude deadlier than the flu, according to all studies I know of that attempt to estimate the Infection Fatality Ratio. The IFR is estimated between 0.5% and 1.6% which is between 5x and 16x deadlier than the flu (0.
652.
by mrb 7y ago
Yes it is. 0.5-1.6% means the coronavirus is 5x-16x deadlier than the flu (0.1%) So the fatality, at the upper end of the range, would be comparable to the 1918 pandemic flu (estimated at 2.04%: https://institutefordiseasemodel
653.
by mrb 7y ago
The denominator was estimated by multiple teams: https://news.ycombinator.com/item?id=22405204 That's why we are fairly confident in the approximative CFR range of 0.5%-1.6% See more details in all the studies I link t
654.
by mrb 7y ago
This NEJM article you referenced ( https://www.nejm.org/doi/full/10.1056/NEJMe2002387 ) is a poor estimate of the Case Fatality Ratio. I'm perplexed it passed peer-review (I'm not even an expert, so
655.
by mrb 7y ago
You are exactly correct. With the formulas you described: naïve CFR = deaths / cases resolved CFR = deaths / (deaths + recoveries) It can be shown pretty dramatically in outbreak simulations how more accurate the resolved CFR is t
656.
by mrb 7y ago
Yes, it is a magnitude deadlier than the flu. We now have 4 independent estimates that place the Infection Fatality Ratio at about 0.5% to 1.7% which is between 5x and 17x deadlier than the flu (0.1%): https://twitter.com/zo
657.
by mrb 7y ago
Then why is China shutting down public transports, telling people to stay home, and cancelling public events? Hmm?
658.
by mrb 7y ago
Most look at infection rates among Wuhan evacuees. Since almost all evacuees were tested, it's a good proxy to estimate the city-wide infection rate. See the 4-5 papers linked in my blog post's last section: https://blo
659.
by mrb 7y ago
I will probably be downvoted for saying this, but... The seasonal flu alone kills half a million people every year. So any mild disease that is as contagious and barely twice as fatal will kill 1 million. And that's what's going t
660.
by mrb 7y ago
Assuming the pandemic will happen, people will buy in bulk and cause shortages anyway. So we are better off getting people to buy food sooner rather than later. This gives more time for stores to replenish stocks.

More ›