UK virus-tracing app switches to Google-Apple model
bbc.co.uk
bbc.co.uk
I'm in the UK and I've almost given up on watching the news as it seems every step taken by the government is weeks behind where it should be, and taken in a reactive way, rather than there being any kind of proactive, coordinated response to the crisis. This is welcome, but months behind when it should have happened - plenty of discussion on here said the route they were trying to take wouldn't work for technical reasons, aside from one's of privacy.
By the time this is up and running, it'll be irrelevant.
I wonder if I can have one day delivery on the worlds smallest violin.
That's... new. He got in power by courting the hard right. He nominated hardcore right-wingers to all key positions, and pre-covid was busy implementing a hard-right "no prisoners" approach to Brexit. And now he'd supposedly be leaning left...?
I agree that he wanted to follow Trump in not locking down anything - but he had to change course after meeting resistance from scientists in the civil service, as widely reported, not from "the left". "The left" didn't even have an actual leader for most of the crisis.
I agree he did not want a lockdown, but I think that happened because France said something like “either you lock down the UK or we ban everyone from the UK from entering our country”.
I can see that he is still in “campaign” mode, but I don’t see how his actions are supposed to win over the left so much as to rally his core supporters.
Then again, if your argument is that he can’t tell the difference, I would agree.
(Yes, I’m British, though I don’t live there any more).
His tendencies and character do a good job of enlightening us as to his lack of principles.
If you think that "lockdown" was to appease the left (who tend to be younger and healthier) rather than the core Conservative vote (the overwhelmingly older, less healthy, more-at-risk-from-COVID group), I don't know what to tell you.
I was involved with free laptops for pupils scheme much trumpeted by UK Government in their press briefings. We finally took delivery of 4 (between 250 children!) on Thursday, months after it was promised and 4 days before children started returning to school...
It's an island nation, it should have had a head-and-shoulders advantage over continental Europe with full potential for an outcome closer to New Zealand or Australia rather than the US or Brazil.
What went wrong?
A government was elected to "get Brexit done", despite being patently (and provably) hopeless at everything else. Then "Everything else" suddenly became really important, and here we are.
Mexico made a big stink about things, and some politicians there even turned their noses up at the U.S. (still the country most likely to render aid) during its peak, but are going into a mass outbreak of their own, that is on a steady trajectory to kill as many or more than the U.S. one.
Mexico: ~20k deaths, 1.55e-4/capita
Sure, the future could be anything, but on the basis of the current evidence that you’re living in a glasshouse, I really don’t think USA Republicans should be casting stones any time soon.
https://ourworldindata.org/grapher/daily-covid-deaths-per-mi...
That's not luck, it's called competence.
> U.S. overall rates of COVID death are dramatically lower than the UK and Sweden
Is 10% less than the 2 worst european countries really worth applauding? Also 10% isn't dramatic in my book. Average US death rate for the last 7 days is still 5 times higher than France for example. [Edit: this is bad metric, see comments below, my bad.]
> people still frame it like the President is a bloody-handed murderer who gets off on watching people suffer.
He willingly took decisions that clearly, scientifically cost the lives of many people to win some political points with his base. You can't possibly think that the crisis was handled completly optimally with the infos available at the time, and no other handling could have saved any more lives.
Our outbreak started later, and so did most countermeasures, for reasons that appear to be complete coincidence.
> Is 10% less than the 2 worst european countries really worth applauding? Also 10% isn't dramatic in my book.
Well, I could've compared Spain and Italy as well. Spain's outcomes seem like they may have a lot to do with vitamin D.
> US death rate is still 5 times higher than France for example.
Right now the rolling-average daily rate is about 3.8x higher in the U.S. than it is in France (which has been in a lockdown of a sort that would probably be illegal in the U.S.); but that's after ignoring the massive spike in France, almost a flat month of death rates more than 2x the U.S. peak (which didn't last as long).
https://ourworldindata.org/grapher/daily-covid-deaths-per-mi...
> He willingly took decisions that clearly, scientifically cost the lives of many people to win some political points with his base.
Such as what? The President is not a dictator, he does not get to make most of the major decisions involved in an outbreak like this. Thanks to everyone scrambling to make things work, it seems that the standard of care was never significantly impacted by a lack of supplies, and none of the hospitals in the hotspot exceeded capacity. Furthermore, the states that most resembled the White House's recommendations have had better outcomes across the board.
> You can't possibly think that the crisis was handled completly optimally with the infos available at the time, and no other handling could have saved any more lives.
That is a completely unrelated claim, that nobody is making. Hindsight is always 20-20. I think that it would have been politically infeasible to do what Taiwan did for all of North America.
Exactly what new information was discovered in the ~1-2 weeks gap between US and Canada outbreaks that led Canada to start lockdown very early while the US started very late?
spoiler: none. One government and it's provinces decided to follow common sense/scientific advices, while the other did not.
> Well, I could've compared Spain and Italy as well. Spain's outcomes seem like they may have a lot to do with vitamin D.
You could have and you would have seen that they have successfully bent the curve to almost no new cases unlike the US. Do you think people in Florida and Texas lack vitamin D? If not how come cases are spiking again there?
> Such as what? The President is not a dictator, he does not get to make most of the major decisions involved in an outbreak like this
Such as not spending weeks trying to convince americans that this virus was a fake news from the left? Or such as using his political weight to pressure states into going into lockdown earlier/longer? Or maybe not vehemently refusing to wear a mask to not appear weak, leading to all of it's base doing the same?
> Hindsight is always 20-20.
I did very specifically say "with the infos available at the time".
How are you calculating this? Comparing the death rates for these four countries one would need to alter the numbers significantly to get the US at 5x France and the UK/Sweden at >5x France: https://ourworldindata.org/grapher/total-covid-deaths-per-mi...
Note: I hope this chart loads. The page is giving me a JS exception so I could only see the chart on the main page (https://ourworldindata.org/covid-deaths). If it doesn't load, scroll down the main page to "Total confirmed COVID-19 deaths per million people" and edit the chart to add those four curves.
You are looking at the entire timespan while I said "Average for the last 7 days". But I agree that this metric is not a good indicator of anything given the lag between when the epidemy started in Europe vs US. My mistake on that.
What does this even mean? "No, they aren't hopeless, they just don't want this challenging situation to exist, so naturally they won't handle it well. You can't blame them for that!"
I mean, they try. Compared to the US government the UK government doesn’t feel architected in any way, it’s just entirely wacky tradition. The House of Lords is totally bizarre for example and that makes me question both the “modern” and “democracy” part.
Then again we have the parts of the USG responsible for the CFR (as opposed to the US code which comes from the democratic part) which is kind of the same thing (I can’t remember what they’re called off the top of my head.)
Personally, most of my friends and I decided to quarantine a few weeks before the government decided to ask people to do so.
I disagree about the contract tracing app though, I'm glad for any reason that it's delayed.
There is no earthly reason I can't do my job from home (already have everything set-up from way before covid) but they want people back in the office...to have all meetings done at desk via videochat.
I pushed hard but they are not moving..which is why I will be as soon as I get a remote offer at comparable or better salary, started looking immediately after the last call where I got a hard no.
Government advice is "Should work from home if able to" but it's advice, not mandatory nor binding and there is no protection/law behind it - so pretty much nothing I can do other than quit, I've savings to get me through a year out of work but I don't want to deplete those in the middle of a pandemic when no one knows what the outcome of all this and a possible recurrence will be - so like a lot of people I'm having to balance the economic risk vs the real health risk about whether to go back or not.
My bosses bosses irrational hatred of WFH is going to put a lot of people at risk.
We are so fucked.
They can barely collectively tie their own shoes, yet suddenly they are responsible for managing the biggest peacetime crisis in our history.
We cancelled a bunch of routine treatment and surgery in order to provide surge capacity. We converted existing wards into new ICUs. We had to take unqualified student nurses and convert them to temporary nurses[1], we took doctors from their specialities and got them to cover other work. Finally, we stopped admitting certain groups of people to hospital altogether. Look at the death rate in care homes.
> They got away with 40,000
That number is for people who died after having a positive test. Many people are not being tested, so there are many deaths of people not included in that figure.
At the moment we think the true number is at least 50% higher, so the real number is 60,000.
And there are different ways to count death. If the UK used the same method to count covid-19 deaths as it does to count flu deaths we'd see over 100,000 deaths.
[1] ...and we said we'd pay them, but now they've finished a lot of them are finding that they're not going to be paid.
Total number of COVID-19 associated UK deaths
42,461
Deaths of people who have had a positive test result
There's lots of evidence. See, for example, this:
https://ltccovid.org/2020/04/12/mortality-associated-with-co...
> Data from England illustrates well the importance of paying attention to differences in definitions and methods used to estimate these percentages: the share of all probable COVID-19 deaths in care homes is 27%, whereas the share of probable deaths of care residents is 38%. The share of excess deaths in care homes during the pandemic has been 44% and the share of deaths of care home residents is 52% of all excess deaths. Also, in France, deaths in care homes are 34% of all COVID-19 deaths, whereas deaths of care home residents are 51%.
That's from May. Since then we discharged a bunch of people from hospital to care and nursing homes, without testing whether they had covid-19 or not.
Contact tracing was launched on 28th May [2].
It seems to me that what went wrong was a large number of UK residents were overseas in February/March (sspecially on ski breaks) to Italy, France and other mainland Europe destinations. The Government didn't impose any emergency isolation or quarantine on those arriving, as was the case for those arriving on repatriation flights from Hubei and other Chinese destinations.
Given much of the scientific analysis of the virus points to strains being descended from those in mainland Europe, it seems a key reason the UK wasn't "head and shoulders" ahead was the scale of this international leisure travel to countries that ended up most affected. Had those returning been put into quarantine, there might have been a very different outcome, but it's hard to say for sure.
[1] https://www.bbc.co.uk/news/uk-52774854
[2] https://assets.publishing.service.gov.uk/government/uploads/...
Given how effective arrival quarantines have been elsewhere (TW, NZ, SG, et al) and how long they’ve been in effect it’s frankly insane that the UK took this long to implement it.
The key to me seems to have been that measures were put in place slowly, at the pace ministers move at, rather than the pace experts can make decisions at. Asking everyone who returned to quarantine for 7 days would clearly have helped, with the benefit of hindsight.
I think the issue is that politicians tend to want to avoid imposing what are seen as invasive measures where they feel the risk is low. Their assessment of the level of risk posed by those returning from overseas holidays in February was clearly now, in the cold light of day, very poor.
The sheer number of people overseas at the time probably didn't help matters - some suggestions that as many as 1 million people were overseas at the time [1].
[1] https://www.theguardian.com/world/2020/mar/19/uk-battling-to...
Frankly, it was clearly very poor even then. That they didn't realise that reflects poorly on them.
I can’t imagine “we’re already drowning, no need to worry about that leak” is really a viable public health strategy in this scenario.
(The reason Italy was reporting so few cases seems to be that they weren't testing people who were hospitalized with symptoms unless they'd travelled to China or had contact with someone who had. There were a few other countries like the US which had similar criteria on paper, but they were generally willing to test other patients based on clinical suspicion. Italy apparently managed to genuinely fail to test hospitalized patients right up until the point where, by sheer coincidence, someone was hospitalized with coronavirus after contact with a traveller from China who turned out not to be the source of the infection.)
This is the same basic situation as Germany. Many of our largest outbreaks in March/April were traced back to a single sick barkeeper in an après-ski in Austria.
Politically, locking down borders was not palatable to the government - I don't think we ever closed the border actually, we just made it very impractical for people to travel. Someone was actually able to get to Paris for a haircut a month or so ago.
https://www.euronews.com/2020/05/18/locked-down-brits-in-nee...
Our care home sector was also more vulnerable than most because it (a) consists of individual care homes which are relatively large and (b) is staffed in large part by agency staff who work in multiple homes.
We're not the only country or region which made this mistake of course and this wasn't the only mistake made but the surprising thing is how long it took for the government to recognise and correct this. Part of that is probably human nature ("we have a plan, stick to the plan"), but I suspect part of it is old-fashioned civil service incompetence too (the same incompetence which led to this being the plan to start with).
They're not. What version of "every country" do you have in mind? Most other countries also failed at contact tracing early on. The only country in Europe that had and used effective contact tracing was Germany. Almost everywhere else it has made little or not contribution.
If you look at the epi curves then the UK, France, Spain, and Italy have had basically the same outbreak (France a little less). So there is no need to wonder why the UK is special because it isn't. There is plenty to say about which European countries responded well and which didn't and why even Germany's response was markedly inferior to countries like Vietnam but that has to start from a data driven perspective that doesn't pretend that the UK is particularly special.
/s
(I'm British in case you wondered... sigh...)
Because of protests and massive intervention of experts and digital citizens rights group like the CCC.
Luckily, the government listened in this case. I am somewhat amazed how well different sectors of society can suddenly work together when the situation is serious.
I am not sure why everything has to be compared in terms of all or nothing.
In this case specifically, they knew that they're playing a loosing game since Apple and Google are ultimately the gatekeepers of the gadgets doing the actual tracing.
Thinking that somehow the UK has the individual power to go ahead and reasonably do it's own thing is nothing but arrogant.
ETA: sources in grandchild comment.
I am trusting Kenny Paterson at ETH Zurich on the cost of the Swiss app, since he was involved in developing it: https://twitter.com/kennyog/status/1273612397861842950?s=20
ETA: NYT article also mentions a "secretive procurement process" for the UK app, which sounds a bit worrying.
Sadly this is unsurprising and par for the course for the UK government these days.
Not particularly secretive, just a bit bureaucratic.
Does the G-Cloud procurement website have any details on where the 108M was spent?
[1] Random website I found when googling, grain of salt etc: https://bylinetimes.com/2020/05/08/nhsx-contract-tracing-app...
It's the same strategy non-democratic regimes use ( Trump, China, Russia)
Some of the feedback seems to be frustrating (for those building a system like this) in that it came down to more human factors like people not wanting to receive this kind of news via a message on their phone, and it scaring them.
I think the constant delaying to the app is a real problem, but it does appear to have become a target of pot-shots from all sides of the political spectrum. Despite this, some of the innovation that went into the app seemed to be quite impressive, and I think it's a shame we won't get to see the outcomes of a "well-thought-out" centralised app compared to the decentralised ones. My concern with the decentralised apps is that they are reliant on testing, testing means latency, and given the incubation period of covid, it seems people will be at their most infectious, spreading it to the next generation of recipients, at the time they are notified they may have been exposed.
At least in my view, getting notifications out ASAP is most important given the pre-symptomatic spreading, and the centralised approach seems to both aid that, and help get people able to stop isolating if it turns out they weren't exposed to Covid (i.e. they and others exposed to same person don't have symptoms in 5 days, thus suggesting a false alarm). There was more to the centralised app than purely "centralised vs decentralised", but there's clearly other issues in the approach to handling covid.
Also, only speaking for myself here: The morning the German App was released, I installed it without hesitation or any trace of doubt. It's now running on my [only] phone, which I have 24/7 with me. So within its limits, the App is rather effective. If it was centralized,... I probably would have hesitated a lot, maybe even skipped it or installed it on my old phone (with the GPS disabled via root), and only turn on that phone on demand. What I want to say: Acceptance/wide-spread use is rather important. Of course the gov can [try to] mandate the app to be used (does the UK do that? idk), but then they're opening another can of worms because they're discriminating against poorer people/demographics who might have older phones with poor battery, limited memory/RAM/CPU, or no smartphones at all (e.g. cheap dumbphones).
Full disclosure: I'm sometimes checking with BeaconScope if the people around me are generally using the app. It's not as many as one would like, but it's slowly ramping up.
The UK has no plan to make the app "mandatory", and they are well aware of the demographic issues that would cause, especially with large areas of elderly populations. That's why there's such a focus on the traditional approach to contact tracing.
RE decentralised vs centralised, the trade-off isn't so simple though in my view - the decentralised app requires the creation of a list of the "infected", which is synced to everyone. That means those who get infected suffer a fairly large "loss" of privacy compared to those not infected. The centralised approach "averages" the two by not broadcasting an infected list. It isn't hugely difficult to use time-based correlation with a calendar or diary to work out if a given person was infected etc, if you have access to the list and database (which we should assume in a threat model people have). The question is if this is a design goal or not.
In addition to the app is traditional "public health authority" contact tracing. That in itself is very privacy-invasive though, and I think people are ignoring it in favour of talking about the app - collecting names and contact info of everyone someone was in contact with, storing this into an IT system, then contacting them etc. Data retention concerns etc and privacy risks here seem significant, yet there has been more focus on the apps than on this traditional approach.
I don't think it works that way.
Here the requirements on privacy which became the accepted base for the German app:
https://www.ccc.de/en/updates/2020/contact-tracing-requireme...
Looking at the Apple + Google spec though, this is happening - see https://covid19-static.cdn-apple.com/applications/covid19/cu...
I can't really see a way to avoid creating an "infected tokens" list, if you want a decentralised approach where the uninfected are not interacting with the protocol on an ongoing basis as well.
The UK approach (centralised) was different though, as it was based on the infected person telling the server "who" (based on the tokens) it saw, and they take the steps to notify those users by decrypting the tokens etc, and working out who needs notified.
In the decentralised apps, your phone squawks a changing "identifier", everyone stores the squawks they hear, and periodically checks their list of squawks heard to see if any are flagged on the infected list.
So it's not the case that privacy was never considered by NHSX, just that they wanted to do things that you cannot do without some degree of centralisation. Most importantly they warn you when a contact develops symptoms and then warn you again if they test positive. That would help get at least some people reducing their contacts earlier.
Apparently despite some very clever design on their Bluetooth system they never got it to work well enough.
Have you seen the news about the government terminating student nurse paid contracts early? Right when we need them the most and when public support for health workers is at an all time the high, our government decide to screw them over. I just don’t get the sanity of such a move.
https://www.personneltoday.com/hr/student-nurses-see-paid-pl...
The way people behaving already makes clear that the pandemic is over in everyone's heads. You might think that especially then an app would be helpful, but we might be talking about 10% of the population actively using it. They probably didn't get the memo the app is completed. Still, I think this sets an important example of how to develop an app as a public service.
If the German government would have developed such an app without loud input from critics, it would probably look exactly like the one the British tried to create. But having a loud voice around issues of data protection doesn't seem to be bad for software at all.
edit: btw: quite some problems wouldn't have been this large if OS for mobile devices would actually be good. And I don't mean extending Apple/Google "features" for protected memory.
Contact tracing scales extremely poorly. Contact tracing a single case involves a long interview and to the tune of a hundred people to be found and tested; for a disease like COVID you need to catch 60–80% of all infected contacts. Rinse repeat for each newly discovered case. It just wasn't possible in the UK for more than a few hundred cases and in the absence of R-reducing measures, so it was abandoned. And before you make comparisons with other countries, just having a centralised database of addresses makes contact tracing much easier, but it's not a thing in the UK because privacy.
Mass testing requires mass testing capacity, which just wasn't there. And it's not clear how mass testing would make anything better anyway, the policy was to self quarantine on any symptoms regardless of covid status.
Not sure if this sudden outbreak of common sense was due to either:
a) technical difficulties of doing something hacky with iOS to try and make the background bluetooth approach work or implementing the backend,
b) something else usurping the need for the data mining (e.g. perhaps the much-trumpeted track-and-trace system where people self-report/voluntarily tell the government who they spend time with)
c) actually caring about people's privacy.
Probably a combination of a + b I expect. c feels like a longshot.
e) speeding up development, because of a) and d), which also means making it cheaper and less reliant on a single supplier
f) the new guy in charge has some Apple contacts he might be able to leverage to, again, speed up things.
Win for privacy.
"I think the people in this country have had enough of experts"
What are you trying to say?
I know that the app used a kind of hack to allow for Bluetooth during sleep (something missing from the Aus app) - maybe they were told that was going to be patched out?
I really think this must be related to it not working as intended rather than anything political.
[1]: https://www.independent.co.uk/news/uk/politics/coronavirus-c...
A crypto coin project from a few years ago has a similar feel to the code. Nothing is going on.
The symptoms were changed over the last few weeks to officially accept that loss of taste/smell was a defining characteristic. The app could have been updated to include this updated self diagnosis information. That would have at least indicated progress. But not even these basic updates have happened.
Pull requests are few and far between. The ones that exist are not dealt with. Again, a sign of a project far from active.
With software we all know that you can do stuff client or server side to a certain extent. We don't know how much of the server side of this app is built. I suspect they put a minimum effort app together just because. There is no evidence of it hooking up to call centres and contact tracing. The people in the call centres are allegedly idle.
The public mood has changed and people are using their common sense rather than following government guidance to the letter. The opportunity for everyone to be running this app has gone. Only a change such as a requirement to have the app if using public transport or entering shops can make it an app people want to use.
One concern I do have with this approach is that the decentralised system inherently can only model the "risk-to" an individual from an epidemiological perspective. It can't model the "risk-from" side of things. To avoid Sybil type attacks, you need to use centrally issued (and thus verifiable) test results to notify others.
Unless tests are administered very promptly, with results immediately available, it seems difficult to make this work in the decentralised approach - with a mean and median "time to symptoms" of around 5.1 days [1], and someone at their most infectious for the 1 to 2 days prior to experiencing symptoms, it seems decentralised apps will struggle.
If A is infected on day 0, then on day 3 they may infect B and C. On day 5, A experiences symptoms, goes online, and requests a test, saying they have symptoms. They realistically would get a test the following day. and the result the day after that. That would be day 7, and at that point, B and C might be notified by the app. The issue is that B and C are now on day 4 of their own infections, the second day of their most infectious period, walking around unaware. They find out the day before they would themselves (on average) experience symptoms.
Clearly the virus is not as deterministic in behaviour as in this simple model, but the above seems to suggest to me that to get any real value out of this (i.e. preventing the next generation of infection), you need to either be able to respond to symptoms reports, or significantly expedite the testing process. In the above situation, a same-day test-and-result would enable B and C to be notified on day 2 or early on day 3 of their own infection, potentially having some real impact. But to achieve this, the latency of testing would need to drop hugely.
[1] https://hub.jhu.edu/2020/03/09/coronavirus-incubation-period...
To my mind, this is a simple matter of reducing the time the "next generation" of infected people are in the community before being notified. If that can be below 3 days, I think the disease can be managed based on the law of averages. If not, it seems it will continue to spread.
That's why, absent a national "Amazon Prime Now" style <= 2 hour delivery window for tests (and some way to verify the results of these tests for input to a decentralised app), I don't see this working.
The app will only really be needed/helpful in cases where you don't know people you are in close or prolonged proximity to, i.e. shops and public transport. In these cases, it makes sense to notify as early as possible, even at the risk of some false positives for those not actually infected, but who were close. The question is how many FPs there will be, but this also happens with human-based contact tracing too.
[1] https://www.theguardian.com/world/2020/may/21/uk-coronavirus...
I don't think the government deserves the credit they have received for this. The "hack" was well understood and documented before this App was created, and it has been implemented poorly [1]. [1] https://github.com/nhsx/COVID-19-app-Android-BETA/issues/10
It wasn't that clever, it didn't work: https://github.com/nhsx/COVID-19-app-iOS-BETA/issues/2
Irish researchers precluded this approach from being effective though -- Bluetooth as a technology is just not reliable enough for this specific use case. https://www.irishtimes.com/news/ireland/irish-news/covid-19-...
https://github.com/corona-warn-app/cwa-app-android/issues/47...
I now live in Finland, with a Finnish address and bank card and everything and I still can't get the damn thing to change. I refuse to accept that it is a technical impossibility to have a UK bank app and a Finnish one (for example) on the same phone.
I do have apps from both stores installed on my phone now (after a lot of trial and error), but believe there is a special place in hell _anyway_ for companies providing services like delivery who do not add their apps to stores globally (looking at you, Doordash). Do they think a tourist or business traveller never wants to order lunch to their hotel?
Sorry if I'm sounding naive, but people are going to start traveling again, so limiting a tracing app to a country's borders seems arbitrarily inefficient to say the least.
There are lots of different legal requirements, and the app points towards the local testing regime for example which is different in every country.
The Apple and Google roadmap already discusses that future items include more interoperability to enable international travel, it's just not a v1 feature.
Palantir of all people shouldn't have been anywhere near the NHS in the first place.
This is the biggest question for me. I’m interested in the actual public health usefulness.
I also think they should have developed two apps in parallel from the beginning. It would hardly have cost much more when you consider the benefits and it would mean that the decision to move from one to the other could be made immediately. (I know they've been working on a Google/Apple one but not sure how much work has gone into it.)
I suspect that they concluded that the whole system won't be ready soon enough (ideally it should have been deployed by now) and so they are switching to something they can deploy faster.
In any case I suspect that these apps will make very little difference because they will come too late, not be used widely enough, and provide too little actionable data.
We cannot avoid most people contracting this virus so why are people keeping running totals of deaths as a score card? I guess we like the idea that we could have stopped it and it gives us a sense of control and something to rage against.
The UK government is estimating that 1/1000 people have had it already. If true that means our final death toll could be in the millions right? we don't talk about that thou, we just say, lockdown should have happened a week earlier! We could have saved 2000 lives. Ok, is that 2000 lives OVERALL, like final figures or is it just 2000 lives by this point?
And the idea that it's just the UK government... 1/6 people said they'd ignore the governments advise on face coverings on public transport. Basically the UK is a load of raging, entitled children for the most part, we don't want to sacrifice anything and we want other people to fix things for us... Cos it's "their job"
South Korea is one of the poster children for "manage to stop the virus entering your country". They have been aggressively responding to each and every case because they are able to identify each case that enters the country. Even then, they are still not fully "reopened".
I agree that neither has reached the level of "destroying international travel and your economy", but let's not understate the impact that their measures are having either.
[0]: https://assets.publishing.service.gov.uk/government/uploads/...
> "make the app compatible with other countries' counterparts, which are based on the same system "
So if multiple countries are using the Apple-Google system, we'll have an international system that anyone can use? I guess this bodes well for future international travel over the next 12/18 months (... or until the vaccines etc are ready) since we'd still get the benefits if we are in compatible countries.
It would be difficult to see how that would ever happen for an app not using the APIs.
Also: " the European Commission said that France - which had adopted a centralised app - would face challenges in this regard."
I think we will inevitably see most developed countries falling in line. On the other hand, I expect the likes of Hungary to go "centralized".
(Ill informed and barely remembered fragments that somehow Palantir / Cummings / Satan's niece was involved in this app are nonsense. Palantir is doing some internal data vis work and both apps are being built by the same team anyway)
The first thing to consider is whether this has led to a real delay in roll-out. We don't know that so far, but we do know that they actually started work on an app based on the Google-Apple API more than a month ago and very few countries have done a full roll-out of an app based on that API yet (and those quite recently). We also know that these approaches work better after "full lockdown" because during the lockdown contacts are already massively suppressed. If it hasn't led to a delay, then the inevitable wailing and gnashing of teeth will be completely irrelevant. Optionality is good, the more things tried the better.
The second is to acknowledge that one of the issues with the (very good idea) of partially or wholly anonymous contact tracing is that apparently people don't like receiving this news from an app interface. That makes an app which interfaces which the rest of the contact tracing system like the one that NHSX originally built a better bet.
The third is to realise that there is as yet no evidence that these solutions work at all. We know that fully centralised contact tracing (which includes a geolocation tracking app) works in Asia. Nobody knows if either the fully decentralised DP-3T system developed by Apple/Google or the partially decentralised PEPP-PT/Robert system developed by NHSX will works.
The fourth is that if you can make it work around phone OS restrictions, there are very real advantages to the NHSX original design which simply cannot be replicated with a fully decentralised system.
There are two which are particularly noteworthy.
First, you can extract aggregated local infection data from the NHSX system which can drive granular local lockdowns.
Second, rather more significantly, because of the central control element, you can revoke notifications. That's important because revokeable notifications allows you to take the risk of allowing some notifications based on symptom onset before a test has been conducted. Since the peak transmission window is around symptom onset and the incubation period is short, notifying early has very significant advantages. People will be irritated if they get a notification which is revoked 24 hours later and you do have to consider the behavioural elements but it is a useful capability to have.
You can also do exposure cohorting. If person A had 10 contacts in a three hour period and on day 8 after exposure none have developed symptoms, you may consider releasing them all from quarantine early. This requires developing quite a sophisticated risk model and carefully weighing up the tradeoffs - you may decide not to use it - but it is an extremely powerful thing to be able to do.
It has never been the case that the Google/Apple solution is "obviously better". Rather there are complex tradeoffs to be considered between different solutions. It would be one thing if we had data to show that the Google/Apple supported solution worked brilliantly. That would mean that the additional features the NHSX design has are not required and therefore the privacy tradeoff is not worth it. We don't though, we know that the a fully centralised panopticon nightmare system works in South Korea and China and we know that we don't want that.
(edit: obviously if they can't make their own app because of Bluetooth issues then the whole game ends and they have to switch)
(edit 2: it was apparently due to technical issues that they were not able to resolve. In denser areas the keepalive mechanism they built works well and keeps most iphones and androids visible for long periods of time but if you go through a period of not being around other phones, too many will drop out. Note that Bluetooth is an inherently challenging way of mapping distance and that this applies to any app which uses it)
That said, my suspicion is that they have good track records of working with open web stacks, and were somewhat less familiar with the realities of working with closed app platforms that relate to hardware integration, and that's where they came unstuck.
If you mean that private companies developing for the public sector in shady deals have a terrible history, I'd agree.