Discover and read the best of Twitter Threads about #nhsxCOVID19app
Most recents (9)
Apropos of @thesundaytimes article, a few thoughts: firstly, as I tweeted 2 days ago 👇 & in stark contrast to the #transparency around other tech/data intitiatives - *cough* @NHSEngland #DataStore *cough* - the team published a blog on what it was doing:
healthtech.blog.gov.uk/2020/10/29/how…
healthtech.blog.gov.uk/2020/10/29/how…
In this blog, the app team gave an explanation of the latest changes it had made - including a reduction in the #RiskThreshold, which took account of TWO factors: the inclusion of a measure of #infectiousness in the API *and* a new #distance algorithm...
healthtech.blog.gov.uk/2020/10/29/how…
healthtech.blog.gov.uk/2020/10/29/how…
In case folks aren't aware, @Google & @Apple's #ExposureNotification API is quite frequently updated. As the #Android developer pages (and #nhsxCOVID18app's blog) say, the #infectiousness field was only added at v1.6; with functionality changing at 1.7 👇
developers.google.com/android/exposu…
developers.google.com/android/exposu…
So the #nhsxCOVID19app currently cannot handle a positive test result from 32.2% of labs, i.e. the whole of pillar 1:
gov.uk/government/pub…
🤦🏻‍♂️🤦🏻‍♂️🤦🏻‍♂️
gov.uk/government/pub…
🤦🏻‍♂️🤦🏻‍♂️🤦🏻‍♂️
Oh, and the app cannot currently handle a #negative test result (e.g. to shut off the #SelfIsolation countdown) from any lab at all!
bbc.co.uk/news/technolog…
Good that @DHSCgovuk will "update" this, but when? And why didn't it do so when we pointed this out before launch?
bbc.co.uk/news/technolog…
Good that @DHSCgovuk will "update" this, but when? And why didn't it do so when we pointed this out before launch?
The sooner all parts of government (and Government too) recognise that some of us aren't *just* "#privacy advocates" - and that we understand many aspects of #digital / #technosocial #systems as well as (if not better than) they do, the better things might be.
Helpful #nhsxCOVID19app explainer [Thread] 👇 from @hadleybeeman.
For more technical background, see github.com/nhsx/covid19-a… for, e.g. system architecture (v3) and (predominantly @AWS) infrastructure.
Of course, we're still waiting for the results of the trials...
For more technical background, see github.com/nhsx/covid19-a… for, e.g. system architecture (v3) and (predominantly @AWS) infrastructure.
Of course, we're still waiting for the results of the trials...
...and, as the BBC's write-up concludes, evidence that #ExposureNotification works in practice is "scant" - in ALL countries attempting it, not just the UK:
bbc.co.uk/news/technolog…
Efficacy (as in NZ) will most likely come from #QRcheckin, which could've been done months ago.
bbc.co.uk/news/technolog…
Efficacy (as in NZ) will most likely come from #QRcheckin, which could've been done months ago.
Bottom line, will I be downloading it? Already have - as I did with the other versions, to see what they did.
Will I be using it? Probably. Not least to avoid having to hand over personal details; combined with regular data deletes/resets.
Will I trust it to keep me safe? No...
Will I be using it? Probably. Not least to avoid having to hand over personal details; combined with regular data deletes/resets.
Will I trust it to keep me safe? No...
So not so much a #ContactTracing app as a feature-creeping Christmas tree?
One thing about tech that's successful is that it is ultimately successful because it performs the core function it (cl)aims to provide well.
#efficacious ≠#effectiveness
One thing about tech that's successful is that it is ultimately successful because it performs the core function it (cl)aims to provide well.
#efficacious ≠#effectiveness
In case you're wondering, "Detection rate" refers to detecting other #phones with the app installed, not the #virus:
The #FalseNegative rate is the amount of times the app incorrectly reports a contact as 'non-infectious' when in fact it is / could be...
The #FalseNegative rate is the amount of times the app incorrectly reports a contact as 'non-infectious' when in fact it is / could be...
...and the #FalsePositive rate is how often the app incorrectly reports a contact as 'infectious' when it's not, i.e. a #FalseAlarm.
Basically, what these #ErrorRates tell us is that #Bluetooth ranging is HARD!
Would you trust a #SmokeAlarm that went off wrongly nearly half...
Basically, what these #ErrorRates tell us is that #Bluetooth ranging is HARD!
Would you trust a #SmokeAlarm that went off wrongly nearly half...
Fascinating piece by @rowlsmanthorpe on the beleaguered #nhsxCovid19app:
news.sky.com/story/coronavi…
Insiders pretty much nail the two core problems: firstly, lack of clarity on what the app was supposed to be & do, and secondly 👇
news.sky.com/story/coronavi…
Insiders pretty much nail the two core problems: firstly, lack of clarity on what the app was supposed to be & do, and secondly 👇
To those attuned to internal cultures, there are some amusing echoes of @ukhomeoffice: @MattHancock's "4%" tirade, reminiscent of @sajidjavid's insistence that @Apple should change iOS to suit HO's #SettledStatus scheme, and some wicked interorganisational chess 👇
The only bit that spoils the piece for me is this 👇
If you've been following its development and read the contracts & DPIA, you'll know @NHSEngland (not @NHSX) is controller of the #COVID19 #DataStore, which sits on @Microsoft @Azure.
@PalantirTech doesn't "run" anything...
If you've been following its development and read the contracts & DPIA, you'll know @NHSEngland (not @NHSX) is controller of the #COVID19 #DataStore, which sits on @Microsoft @Azure.
@PalantirTech doesn't "run" anything...
And so it begins...
"If you won't put your details on this list, you ain't coming in" is a slippery slope - via "#ImmunityPassports"? - to the (digital) #PapersPlease culture that the #DatabaseState proponents have always wanted.
Muse on that while you sup your pint.
"If you won't put your details on this list, you ain't coming in" is a slippery slope - via "#ImmunityPassports"? - to the (digital) #PapersPlease culture that the #DatabaseState proponents have always wanted.
Muse on that while you sup your pint.
N.B. It's also a wicked marketing set-up for when/if the #nhsxCovid19app launches: "Oh, you're #surveilling yourself, madam - please walk right in! No need for that inconvenient list any more..."
What other #freedoms will folk #surrender for a mere 100cm?
What other #freedoms will folk #surrender for a mere 100cm?
Please note, I am not advising against - in fact I endorse! - sensible #clinical & #PublicHealth advice. Problem is, they just crashed into @10DowningStreet's idea of the #economy. (About which we'll be told more shortly, no doubt...)
Guess who loses?
Guess who loses?
Yup. The 'evolution' is fascinating, but I fear anyone who thinks @NHSX has the IN-HOUSE technical capacity to develop an app like this sorely misunderstands what #DHSX actually is, e.g. it had to hire @faculty_ai to design, staff & set the strategic direction for its "#AILab"...
...as the first 6 bullet points on p7 of this contract show: cdn-prod.opendemocracy.net/media/document…
According to recent reports, the cost of the #nhsxCovid19app's third party development also grew quite a bit:
though, as @EerkeBoiten says, one must be careful not to...
According to recent reports, the cost of the #nhsxCovid19app's third party development also grew quite a bit:
though, as @EerkeBoiten says, one must be careful not to...
...conflate either (a) what each project is, (b) which companies are working on each, albeit some on several, and (c) who is the notional* data controller.
__
*"Notional", as controllership's down to who determines purpose & MEANS of processing - not mere contractual assignation
__
*"Notional", as controllership's down to who determines purpose & MEANS of processing - not mere contractual assignation
"#Power" is certainly one framing; #incompetence & #hubris two others.
Like Chris, I reckon @NHSX has (finally) made the right decision - one that was quite obvious both pre-#GApple
medconfidential.org/2020/apps-for-… & after medconfidential.org/2020/gchq-and-…
... & the damn thing still has to WORK!
Like Chris, I reckon @NHSX has (finally) made the right decision - one that was quite obvious both pre-#GApple
medconfidential.org/2020/apps-for-… & after medconfidential.org/2020/gchq-and-…
... & the damn thing still has to WORK!
See also:
And this:
...though I think "lessons learned" may be a bit precipitant / generous to the Gov't in the circumstances. There's zero sign of (2) as yet, and active *avoidance* of both (5) & (6).
...though I think "lessons learned" may be a bit precipitant / generous to the Gov't in the circumstances. There's zero sign of (2) as yet, and active *avoidance* of both (5) & (6).
It really isn't. As Peter points out in his [Thread] @CDEIUK avoids talking about what UK Gov't is planning to do, and carefully chooses examples that skirt the very real #discrimination issues - Yellow Fever? How about HIV status? And effectively mandatory #biometricID?!
The not-very-subtle "We can't do nothing, #ImmunityCertificates are the only option" framing is redolent of #Whitehall & #WestminsterBubble #groupthink, barely tempered by the admission we don't know enough about #COVID19 #immunity to reasonably consider such things. Meanwhile...
...evidence is emerging that some people don't develop #antibodies, even after having caught the #virus:
news.sky.com/story/coronavi…
At this point, #ImmunityPassports have even LESS chance of working / being useful than the much over-hyped #nhsxCovid19app - but with far greater...
news.sky.com/story/coronavi…
At this point, #ImmunityPassports have even LESS chance of working / being useful than the much over-hyped #nhsxCovid19app - but with far greater...
