Neither site is clear where the testing capacity goal comes from. For NY, it is 30 per 1k (3%). For Bay Area, it is 200 per 100k (.2%). If my math is right, that's a significant discrepancy.
Actually, I see NY's goal is monthly. So I guess that is more like 30 per 1k per 30 days (.1%).
There some discussion of necessary testing rates in this NY magazine article:
It looks like .2% is at the low end of most ready-to-reopen requirements. In any event, it looks we're nowhere near a test for anyone who wants one.
Also, does anyone know of a good chart that graphs daily new cases vs new tests? It seems like that ratio would be significant, especially when comparing case rates between regions.
As far as I know, it's still the case that no reasonably-large country is doing much more coronavirus testing than the US right now. The UK is probably just about over the 30 tests per 1k per 30 days mark averaged over a week as of our last testing increase, Germany might still be slightly below it but I haven't seen recent figures from them and they seemed to have stalled out increasing testing, and other countries are similar.
FWIW my province of 4.4 million is stalled on tests being run because it's run out of additional symptomatic people to test at around 30 tests per 1k per 30 days, even though testing capacity is about double that, and still increasing.
1. You can't actually randomly choose people, so it's being done on a fist-come, first-serve basis.
2. Testing all asymptomatic people on any regular basis would take orders of magnitude more tests, so it's not useful as a containment measure, just as a data-collection measure.
That ranking's not terribly informative. A lot of the countries heading it up are smaller ones since most countries are drawing from the same global pool of testing consumables, meaning that there's not much relationship between the difficulty of carrying out a particular number of tests and the size of the country, it's total tests done which doesn't tell you much about current testing rate, and large developed Western countries are close enough in testing rates that an actual ranking based on them probably wouldn't be very stable anyway.
The problem with talking about per capita testing numbers is that having a bigger population means you need a proportionally larger number of tests to administer to reach the same numbers as smaller countries. The will to test and the money to do so aren’t any issue. The problem is that we can’t just snap our fingers and make tens of millions of tests appear out of thin air. I think this point gets lost on a lot of people when they look at the corona scoreboard (likewise, people fixate on the large number of deaths in the US and ignore that we’re doing better than several European nations on a deaths per million basis).
> The problem is that we can’t just snap our fingers and make tens of millions of tests appear out of thin air.
You are correct - but we (or rather the government) can snap it's fingers and compel some private companies that are well positioned to be able to manufacture tests to do so at a higher priority than other business concerns - or we (again the government) can issue generous contracts for testing supplies that guarantees payment to private companies manufacturing them even if the original quota of tests requested by the government is above the level we end up needing.
I think that large countries actually have an advantage here - a small country might not have any internal industry that'd be capable of manufacturing tests without heavy retooling - or that industry might be so small and specialized that scaling it up is infeasible. But even in that sort of a situation they can use market based solutions to bid on tests in a manner that motivates private companies in other countries to feel confident committing to test production - and that's only needed if there isn't any sort of altruistic world-banding-together-to-fight-the-issue effort.
> can snap it's fingers and compel some private companies that are well positioned to be able to manufacture tests to do so at a higher priority than other business concerns
And we’ve done that with things like masks and ventilators. But there is a key limiting factor with the tests, the reagant supply. Again, can’t just snap our fingers and make it happen faster. This isn’t AWS, we can’t just spin up more supply in an instant.
America does have just as good of an ability to secure reagent supplies as any other countries though, excluding the ones that produce it domestically. The exception time this is that different testing machines require different reagents.
It’s not like every country is getting a ration of the same number of units - so it shouldn’t be harder for large countries. Again this is likely an issue for smaller countries as they have to compete with America for a share of the reagent supply.
Sure, but it likely has ten times the infrastructure to manufacture and deliver them.
The US is, in many ways, a bunch of ~30M person countries, akin to the European Union. Freedom of movement, no customs between them, powerful state-level governments, etc.
> Sure, but it likely has ten times the infrastructure to manufacture and deliver them.
Yes, in theory, but we're talking about trying to produce 10x of something limited by a resource in short supply the world over, with every country's government vying for their own cut. Instead of "reagent", pretend I told you the limiting factor was "Unobtanium." Hey no sweat, just find 10x of it! See how silly that line of reasoning is? Again, to put it in terms the HN crowd might understand, the production of physical goods isn't like spinning up servers in AWS.
Or are you just understand the impression that the US, who has produced by far the greatest number of tests (in absolute numbers) of any country in the world, just doesn't have the desire to obtain greater and greater numbers of tests?
The claim was that it’s harder for the USA to get the required levels of unobtanium than a smaller country because it needs more.
In reality the USA has more purchasing leverage because it can exploit (and is exploiting) geopolitics, and can obtain at least as much reagent/unobtanium per capita as other countries right now.
To put it in terms you might understand, the USA has a better ability to source reagent than you do - see if you as an individual are able to procure even 1 litre of the stuff!
> Or are you just understand the impression that the US, who has produced by far the greatest number of tests (in absolute numbers) of any country in the world, just doesn't have the desire to obtain greater and greater numbers of tests?
Yes. (I've answered this in more detail in the other comment thread, but Trump has repeatedly complained that testing = detected cases = bad look for him.)
There's no indication the Trump administration has invoked the Defense Production Act to ramp up reagent production. They'd be wise to do so, if it's a genuine bottleneck.
> "They would like to have the people come off. I’d rather have the people stay [on the ship]. But I’d go with them. I told them to make the final decision. I would rather — because I like the numbers being where they are. I don’t need to have the numbers double because of one ship that wasn’t our fault."
> "So, in a way, by doing all of this testing, we make ourselves look bad."
He's not wrong though, look at the rest of the quote in context:
> "So the media likes to say we have the most cases, but we do, by far, the most testing. If we did very little testing, we wouldn't have the most cases. So, in a way, by doing all of this testing, we make ourselves look bad,"
The correctly identified problem with increased testing is that lots of super geniuses, who apparently can't look past "big number is biggest" w.r.t. corona deaths, try to tie "most deaths" with "worst handling of the situation", completely ignoring, as I pointed out, that the US is doing better on a deaths/million basis than several European countries. Countries that supposedly have superior healthcare systems, healthier populations, and leaders not named Trump. Buuuuttt, no one wants to tie context like that to any of the numbers, so he's absolutely right that increased testing => increased reporting of infection => increased whining about how bad the country is doing.
And of course, this is all ignoring that it's not his call to make as far as how many tests are performed, otherwise we wouldn't be leading the world in absolute numbers of tests administered by a staggering degree. No, we'd be content to test nobody at all and brag about how no one is infected.
But none of that matters. You have x number of people, you need y number of tests. Just because it's harder to do doesn't mean you need to do it any less.
Yes it does, it’s proportionally more difficult. If you want a country of 300M to have as many tests per capita as a country of 30M, you need _ten times as many tests_.
Question for HN experts:
what would be the best approach for object-detection when the object I am trying to detect is a specific image (not a class of objects) printed on a 2D surface (therefore it could be partially folded / curved, but still more of a distorted 2D object than a complex 3D one)?
Pyimagesearch does a cool method where they enable template matching to work at different scales. It ignores color information (by using edge detection), which could be a critical feature if the object has specific color:
