Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The EU Commission accepts high error rates when checking chats

netzpolitik.org

Post reply on HN
Print view View original

411–420 of 443 posts

Re: The EU Commission accepts high error rates when checking chats

#411

Post by em-bee »

easytiger wrote: post #409
ars wrote: post #408

Earlier quoted context omitted.

Recording the street is the most important part of recording for crime detection. If someone knocks off the side mirror of my car by driving too close, I shouldn't be able to record that? I shouldn't be able to capture the license plate so I know who it is? You do get that you are publicly visible when walking down the street, right? People are visible in their gardens too, but I can at least understand making it ill…

Or go live in Africa where people case the street before they come to rob/murder you. It's a big deterrent to have CCTV covering property perimeter

it's a band-aid to a different, more serious problem. you both would have a point if the crime rate in germany or austria was higher due to not allowing surveillance in public spaces.

but thankfully they have other deterrents to crime that allow everyone to have more privacy in public spaces.

Re: The EU Commission accepts high error rates when checking chats

#412

Post by vidarh »

AnonCoward42 wrote: post #334
GekkePrutser wrote: post #288

Earlier quoted context omitted.

Big cities are always a lot more progressive. But a christian-conservative like Merkel wouldn't have come to power if it wasn't for the more rural masses. They must have enough votes to influence policy in such a big way. And those are the ones that are the most susceptible to 'for the children!!!' type privacy invasion. At least in my experience here in the Netherlands, this is usually the angle those measures are p…

Merkel is not christian-conservative except for the label of the party itself. She hit most of the progressive boxes really, especially "no nations, no borders" and the climate sale of indulgence. I also wanna add that Merkel was elected by big city voters as well and not just the (framed stupid and wrong) "rural masses". > They must have enough votes to influence policy in such a big way. And those are the ones that…

This sounds like you're coming from a very US viewpoint. By European standards she is firmly conservative.

Re: The EU Commission accepts high error rates when checking chats

#413

Post by vidarh »

hammyhavoc wrote: post #320
GekkePrutser wrote: post #288

Earlier quoted context omitted.

Big cities are always a lot more progressive. But a christian-conservative like Merkel wouldn't have come to power if it wasn't for the more rural masses. They must have enough votes to influence policy in such a big way. And those are the ones that are the most susceptible to 'for the children!!!' type privacy invasion. At least in my experience here in the Netherlands, this is usually the angle those measures are p…

That's very interesting re Hetzner. When I signed up Hetzner about two weeks ago, I wasn't asked for any kind of ID, in fact I was quite surprised how little friction there was. Experience hasn't been good though, so you haven't missed out on much. Downtime and connectivity issues have meant hours of time wasted on an issue that wasn't even at our end. Will check out Scaleway!

That's odd, I have never experienced downtime and connectivity issues there, with constant use for over ten years.

Re: The EU Commission accepts high error rates when checking chats

#414

Post by vidarh »

ars wrote: post #408
em-bee wrote: post #391

Earlier quoted context omitted.

if you happen to cover a portion of the street leading up to your property you are breaking the law in germany too. why is that absurd? you can monitor your own property, but don't record me when i am walking along the street please.

Recording the street is the most important part of recording for crime detection. If someone knocks off the side mirror of my car by driving too close, I shouldn't be able to record that? I shouldn't be able to capture the license plate so I know who it is? You do get that you are publicly visible when walking down the street, right? People are visible in their gardens too, but I can at least understand making it ill…

Regulation of CCTV generally takes the approach that there's a substantial difference between you standing there with a camera and CCTV because you standing there with a camera produces a substantially different level of signal that people are being observed.

Your right to take the footage is in competition with peoples right to a reasonable level of privacy even in public spaces, as there is a huge difference between being seen by someone, being filmed by someone when they happen to be there, and being constantly filmed by automated systems that may or may not be appropriately controlled to ensure images are not spread.

The approach taken tends to vary between outlawing it or allowing it with restrictions, including making you subject to retention and subject access policies.

E.g. in the UK, which is notorious for the amounts of CCTV, you can cover public areas if necessary, but you then need to be prepared to delete footage of people on request, respond to subject access requests, ensure the footage is kept securely, ensuring the footage is deleted regularly, and restrict access to the footage. You also need to be able to provide - in writing if requested - legitimate reasons for doing so. Here's the ICO's advice page on the subject[1]

[1] https://ico.org.uk/your-data-matters/domestic-cctv-systems-g...

Re: The EU Commission accepts high error rates when checking chats

#415

Post by vidarh »

dane-pgp wrote: post #397
vidarh wrote: post #381

Earlier quoted context omitted.

If the goal is to create an authoritarian dictatorship, then sure, controlling "most people" is usually enough because you can control the press by sending thugs to their offices. But your "next step" is far beyond what even China does. It would kill off any ability for software development. I'm all for being vigilant, but these scenarios are not realistic. As for a general purpose computing device being useful, as l…

> But your "next step" is far beyond what even China does. And yet online services blocking access to non-SecureBoot devices is already the norm for industries ranging from online gaming[0] to fast food![1] Add in the fact that Hollywood wants this (for DRM and blocking torrenting apps) and governments like Australia claiming their laws trump the laws of mathematics[2], and you can almost guarantee that this is going…

> And yet online services blocking access to non-SecureBoot devices is already the norm for industries ranging from online gaming[0] to fast food![1]

Specific online services doing so is very different from a general ISP ban. A general ISP ban is impossible as long as you have an IO channel of any kind, including projection of text or playing sound. See the end of this comment. SecureBoot in itself also does not in any sense stop general purpose computing of unsigned code.

> Software development would have to be sponsored by approved companies, or at the very least you'd need to apply to the government for a "licence to code", with your ID number baked into every app you create.

... and you've just kneecapped your software industry in favour of companies outside of said authoritarian hellhole. Won't happen. The EU has a long history of crazy demand like this being proposed, and they end up dying or getting watered down to nothing because there's nowhere near sufficient support for going as far as you suggest.

> But which app are you going to use to create those sound or image files? I suppose you could create your "illegal" files on an airgapped non-approved device, and transfer them via USB to the approved device,

Missing the point. If you can play and record sound on an approved device, for example in a call, or transfer text, no matter how filtered, you can use that as a channel for an non-approved device. We used to use heavily filtered low-fidelity audio channels to transfer data, via acoustically coupled modems, after all. Any attempt to filter this just reduces to making it seem more plausibly like acceptable material, e.g. encoding it in speech for example. This is not even a hard problem, though data rates would be low. If a channel can transfer language, it can transfer data.

But we're talking a regime more oppressive than China for this to even be relevant. Even in China today, "normal" VPN tech is sufficient, though a hassle.

Re: The EU Commission accepts high error rates when checking chats

#416

Post by easytiger »

em-bee wrote: post #411
easytiger wrote: post #409

Earlier quoted context omitted.

Or go live in Africa where people case the street before they come to rob/murder you. It's a big deterrent to have CCTV covering property perimeter

it's a band-aid to a different, more serious problem. you both would have a point if the crime rate in germany or austria was higher due to not allowing surveillance in public spaces. but thankfully they have other deterrents to crime that allow everyone to have more privacy in public spaces.

> germany or austria was higher due to not allowing surveillance in public spaces.

Have you seen how many stabbings have been happening in germany lately?

Re: The EU Commission accepts high error rates when checking chats

#417

Post by n4r9 »

rob_c wrote: post #398
n4r9 wrote: post #394

Earlier quoted context omitted.

At the time, the "base rate" of vaccination was already up around 90%, so the 60% rate in hospitalised patients actually demonstrated significant protection.

Yes again in the art risk groups this reduced significant symptoms hence why they're underrepresented in global hospitalisation numbers. As I keep saying it's a fantastic preventative treatment. Unfortunately many people are still catching it repeatedly, displaying symptoms repeatedly and the fact that despite 90%+ of people are immunised. With vaccination levels like this for various other diseases transmission woul…

This person was questioning its efffectiveness simply because more people in hospital were vaccinated than weren't. Without considering the base rate. A sufficiently high base rate would lead to this situation even if the vaccine was 99.9% effective.

To address your other point, I believe studies have shown a significant reduction in hospitalisation for young adults who take the vaccine, e.g. https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

> only small (and sometimes non-significant) reductions in the risk of hospitalisation were seen in 10–19-year-olds, with increasingly large reductions seen with age in 20–69-year-olds

https://els-jbs-prod-cdn.jbs.elsevierhealth.com/cms/attachme...

Re: The EU Commission accepts high error rates when checking chats

#418

Post by estebank »

easytiger wrote: post #416
em-bee wrote: post #411

Earlier quoted context omitted.

it's a band-aid to a different, more serious problem. you both would have a point if the crime rate in germany or austria was higher due to not allowing surveillance in public spaces. but thankfully they have other deterrents to crime that allow everyone to have more privacy in public spaces.

> germany or austria was higher due to not allowing surveillance in public spaces. Have you seen how many stabbings have been happening in germany lately?

Germany is somewhat safer than France and 6 times safer than the US, and continues to have a downwards trend.

https://data.worldbank.org/indicator/VC.IHR.PSRC.P5?end=2020...

Re: The EU Commission accepts high error rates when checking chats

#419

Post by easytiger »

estebank wrote: post #418
easytiger wrote: post #416

Earlier quoted context omitted.

> germany or austria was higher due to not allowing surveillance in public spaces. Have you seen how many stabbings have been happening in germany lately?

Germany is somewhat safer than France and 6 times safer than the US, and continues to have a downwards trend. https://data.worldbank.org/indicator/VC.IHR.PSRC.P5?end=2020...

How's that in any way relevant?

Re: The EU Commission accepts high error rates when checking chats

#420

Post by rob_c »

n4r9 wrote: post #417
rob_c wrote: post #398

Earlier quoted context omitted.

Yes again in the art risk groups this reduced significant symptoms hence why they're underrepresented in global hospitalisation numbers. As I keep saying it's a fantastic preventative treatment. Unfortunately many people are still catching it repeatedly, displaying symptoms repeatedly and the fact that despite 90%+ of people are immunised. With vaccination levels like this for various other diseases transmission woul…

This person was questioning its efffectiveness simply because more people in hospital were vaccinated than weren't. Without considering the base rate. A sufficiently high base rate would lead to this situation even if the vaccine was 99.9% effective. To address your other point, I believe studies have shown a significant reduction in hospitalisation for young adults who take the vaccine, e.g. https://www.thelancet.co…

> A sufficiently high base rate would lead to this situation even if the vaccine was 99.9% effective. In principle for these factors alone statistically yes, in reality we know this isn't the case. Calling it a vaccine is calling it broken. Calling it a preventative treatment to improve your outcome is more accurate.

The studies showing reduced admissions are either grasping at statistical straws or a reglorofied statement of "This is a novel contagion in a herd without immunity at which some unknown percentage may be clinically at risk from genetic or environmental factors."

The strongest environmental factors we now _know_ (and had clear statistical evidence for as of July2020!) was vitaminD and a few other well defined risk factors.

Hospitalisation in the UK due to exclusively COVID factors reduced dramatically as the pandemic went on due to less strong concern once almost every a&e had seen the full disease cycle from admission to treatments to outcome. Not to mention most COVID cases at hospital in the UK were either contracted there in the younger age groups or they tested positive but this wasn't the concern at the time of admission.

One of the major concerns in May2020 was that "up to 30% of the working population may be off it at one time if the virus spreads according to China's numbers and the economic cost of that is too high..." In hindsight this statement didn't compare to the damage of being frozen by indecision combined with no global trade for 2 years, so... As for guessing how many would end up in hospital report 9 models said the peak would be over 2million needing a&e at the same time, despite the model not combining medical risk with any mortality factors, this was simply 10% or so of total cases if 20M or so we're ill at the same time.

Again, as a preventative treatment for clinically at risk people (who we did a great job of identifying based on medical history in 2020) we should have never had a 2nd global lockdown, unfortunately the particular group all vote a certain way so...

Post reply on HN