Posts by Cuddles 2337 publicly visible posts • joined 3 Nov 2011
Once again, look at the list of countries actually involved in this investigation. Europol already shares intelligence and conducts joint operations with the USA directly. It's just sheer nonsense to suggest that the UK could therefore have a problem just because it might start making deals with the USA.
"We can only hope that Europol will continue to work with the UK now we're a "third country"."
Of the seven countries named in your quote, two are members of the EU. It seems a little dishonest to conclude that Europol might therefore have a problem working with the UK.
"Right, but what are "real conditions"?"
Real conditions are the conditions that actually exist in the real world. Things get mislabled, pictures aren't always perfect professional quality, different equipment models are used in different places, and so on. You say it seems perfectly reasonable to insist that your system only performs when everything is perfect, but that's the whole point - nothing is ever perfect in the real world. If you want a system to actually be useful, it must be able to cope things not being perfect. Showing your system works with frictionless spherical cows under ideal lab conditions is great, but if it falls apart when the cows are no longer frictionless and occasionally turn out to be horses, it's not actually useful.
"Stephen Odaibo, CEO and founder of Retina-AI Health, told us in a statement he thought the researchers' conclusions were not supported by the study.
"First, the study was a retrospective study based on heterogeneous unstructured data..."
This and the following comments seem to be rather missing the point - the problem is specifically that the AI doesn't work in the real world where it has to face heterogenous unstructered data. If it's only useful when you already have a doctor sorting through all the input to make sure it's acceptable, you can just have the doctor do the whole job and throw out the AI. Either it it needs to work with the data that is actually presented under real consitions, or it simply doesn't work at all.
"I think you have to remember that this is the first global pandemic since the 1919 Flu pandemic."
It's the ninth. Plague, two cholera, three flu, typhus, psittacosis and AIDS. Note that the AIDS pandemic is still very much ongoing, and the most recent flu pandemic was only just over a decade ago. Also note that the first SARS, MERS, and Zika all had global spread and only really avoided being declared pandemics based on semantics, all of them within the last two decades.
People like to remember the Spanish flu and talk about how much things have changed since then. But pandemics aren't occasional rare events, we actually average about one per decade. This isn't the first pandemic with global news. It's not even the first with Twitter. There's nothing different this time; we've watched exactly the same things unfold in the same ways over and over again. Everyone just immediately forgets all about it, and then accuses health officials of fearmongering every time they warn about the latest worrying disease with pandemic potential.
"makes a determination based on the facial features of a user and can adjust its understanding of a user over time. In theory, this makes it able to respond to changes in a user's appearance, such as a new haircut or glasses, while also accounting for the changes that gradually occur over time, like fluctuations in weight."
Bets on how long until someone figures out how to exploit this? The whole point of things like passwords is that you need to have an exact match to a known quantity, otherwise it just doesn't work. Even with biometrics at least the target is known and you need to be close enough to work. Meanwhile the recurring theme of "AI" is that it's a black box that finds matches based on some unknown criteria that inevitably turns out to be exploitable when things outside its learning dataset are encountered. Putting your security in the hands of some sort of machine learning box that makes up new login credentials with no input from you sounds like a seriously terrible idea.
"an eight-core Ryzen 7 4800H processor. It also comes with a discrete GeForce RTX 2060 GPU, which will come in handy for those working in creative fields, as well as those partial to high-end gaming."
I'm not convinced a laptop CPU and the cheapest GPU from the last generation really counts as high-end gaming. As a lightweight all-in-one it will probably do the job OK, but you're certainly not going to making use of that 4K resolution in many games. And if you're not fussed about saving space, you could get a decent gaming PC for the same price.
"Has nobody else spotted that the Tripod's legs come from the bottom truncated cone-shaped part of the machine while the strange thing at the front is coming from a hole on the bottom of the saucer section?"
No it isn't, it's exactly the same as the other three legs, the first articulated part just happens to be mostly in front of the body at the angle it's viewed (well, drawn) from.
List price is not profit. They've actually delivered a little over 300 of them. Count the development costs, supposedly as much as $3 billion, and they're very unlikely to have made any profit at all so far. The extra 400 or so that they haven't been allowed to deliver have been actively costing them money. Thats is, after all, the whole reason to rush it through without proper certification in the first place - plane are expensive to design and build, and don't have a huge prodit margin, so there's always a big drive to reduce cost wherever possible and get things delivered to customers as soon as possible.
An even simpler way to look at it is to ignore all the details and just look at Boeing as a whole. In 2019 they lost over $600 million. In 2020 they weren't able to sell a single one of their shiniest new plane, and in fact had a significantly number of orders withdrawn, and that's before you even start looking at the near collapse of the plane industry due to the pandemic. They've now been fined an additional $2.5 billion. There can really be little question that this is actually a pretty significant hit to their finances.
"GPs get paid per-procedure for all procedures they carry out"
Just to further clear up this point - there is an important difference between a GP as in a person and a GP as in the business of a GP practice. GPs get paid per procedure, yes, but it's the business that is paid, it's not some cash slipped directly into a specific doctor's pocket. So complaining that nurses should get paid instead of GPs is just nonsense, neither nurses or doctors are getting any extra pay here, it's the practice that employs them both that gets the money. What actually happens to that money will be up to the management of the practice and the contracts of their employees. It's entirely possible that unscrupulous partners in some practices will pocket all the money for themselves, but it's also possible that nurses will get most of it since they'll be the ones actually doing the extra work and needing overtime payments.
"Trendforce expects Xiaomi will replace Huawei in third place, behind Samsung and Apple, but ahead of Oppo and Vivo."
Or to put it more accurately, Trendforce expects Xiaomi will replace Huawei in fourth place, behind Samsung, BBK and Apple. Oppo and Vivo, along with several others, are the same company.
"While the Orion has cost over 17bn to date, its a fraction of apollo in 'todays' money the eagle cost...."
In today's (well, 2019's) money, the Apollo program cost about $160 billion. The included the full development, manufacture and launch of the whole rocket, capsule, landers, and everything else involved. It includes the cost of actually getting to the Moon 10 times (and 11 past Earth orbit counting Apollo 13), plus four orbital and three sub-orbital test flights of various bits.
Orion has cost over $21 billion in today's money just for a capsule that has had a single orbital test flight and one abort test. The actual intended launch system has so far cost $19 billion and has never flown. Yes, Apollo was expensive. But at this point you can't compare the costs to Orion because there simply isn't anything to compare it with - $160 billion for a successful program with multiple Moon landings, and $30 billion for a half-finished capsule and a rocket that has never been tested. If Orion ever actually achieves anything, then we can start comparing the cost of getting there.
Indeed. Every time I say that there's no evidence all this targetted advertising actually works, the inevitable response is that everyone wouldn't keep spending money on it if that were true. Here we see yet another case where not only did no-one know if their advertising was working, they didn't even notice that it was being shown to anyone at all. And the article claims 10% of the entire market is fraudulent. Asking whether advertising works is really a bit premature at this point, they have yet to show any evidence they can tell if it's even happening.
"Google argues the lawsuit seeks to dictate how it designs its products and services."
Yes, exactly. The entire point of laws and regulations is to dictate rules that products and services must follow. If you design a product and/or service that in some way goes against what society considers acceptable, you will be forced to change your design. Things like building codes, electrical safety, not filling food with pesticides and razor blades, they're all just dictating things you have to consider when designing your products. Rules regarding monopolies and anti-competitve behaviour are no different, they just address harm that tends to be less tangible and immediate.
Large parts might be putting it a bit optimistically. Australia and South Africa both have some test kit set up on site. Out of the planned 3000 dishes and 130,000 attenae, they have 100 dishes between them, all significantly smaller than the ones that will be used in the actual SKA. It's an important part of the design and prototyping process, and obviously can do plenty of useful science in its own right, but it's a very, very long way from being a large part of the final thing.
"I'll be moving to Protonmail"
Protonmail only allows plus-aliasing which far too many places don't accept as valid email addresses. If you don't care about that then Protonmail is great, but for people who want a bit more flexibility in handling aliases and multiple addresses I'd recommend Fastmail instead.
"I have my email through a paid service. It's a minimal cost"
Absolutely. People have become far too used to having email just there as a free service, so almost any cost is seen as too high. But considering how important email is to almost every part of your life these days, as effectively the sole online identifier, £5 or so a month hardly seems excessive. I pay about £5 per year for my own domain as well, and while that's optional it's still not exactly a huge cost. Hotmail and gmail seemed great when they first appeared, since the alternative for the average user was to be stuck with whatever their ISP provided and the consequent problems if they tried to change, or potentially even just moved house. But we're well past the point where the benefits of those free services outweighs the costs. The trouble is that the cost isn't monetary, so most people don't notice just how big it is.
Simply listing a price per bit also misses out the various less easily measurable and/or comparable factors. Is a 1 Gb/s service that drops out for a few hours every week better value than a 50 Mb/s service that works perfectly all the time? How about a fast service with lots of congestion and throttling preventing you from ever getting the headline number, compared to a slow one that always gives what you pay for or more? What about a service with worse reliability but great customer support, compared to one that is much more reliable but hopeless when something does go wrong?
Of course, I'd be reasonably surprised if the UK isn't lagging on most of these aspects as well. But it would be nice to see some efforts to actually evaluate them. A simplistic price per bit comparison really isn't that useful if you value things like reliability or support more highly than just headline speed.
"Is pretty cool, suspect such a thing could also be achieved by blinking a status light or something."
Yep. Hard drive light - https://www.theregister.com/2017/02/23/hard_drive_light_used_to_exfiltrate_data/
Router light - https://www.theregister.com/2017/06/06/data_exfiltration_with_routers_leds/
Also monitor pixels, LCD brightness, drive noise, power cables, case temperature, and basically any property of a computer that can be in any way controlled or monitored. Every time you see a headline about getting data out of a computer in some seemingly insane way, you can pretty much guarantee Bu-Gurion University is involved.
As for the inevitable whining about these attacks not being practical because it requires access to the machine, that remains just as stupid as ever. Just because an attack requires physical access doesn't mean it's irrelevant because then an attacker could just do anything they want. We even have a variety of names to describe some of the circumstances where physical access is very relevant. It usually involves either compromise of something you trust, as in supply chain attacks, or access for a short time, as in evil maid attacks. In both cases, physical access provides the initial compromise, but the attacker still needs some way to actually do anything afterwards.
That's the entire point of this sort of research. The traditional approach to guarding against attacks like that is to air-gap machines - don't connect to the internet, block off the USB ports, and so on. Even if your supply chain is compromised, it doesn't matter because you never connect to the outside world anyway. What Ben-Gurion keep showing is that there are all kinds of ways to get data out that aren't normally protected against. It doesn't matter that the proof-of-concepts aren't usually especially practical or that most of them are fairly trivial to block once you know about them. If you're paranoid enough to worry about evil maids and supply chains, you also need to be paranoid enough to do more than assume that just because you haven't plugged an internet cable in everything must be secure.
"AI and automation in the workplace risk creating new forms of bias and unfairness"
The whole problem with "AI" is that it doesn't create new anything, it just repeats the problems that were present in the input. It may provide a new excuse for perpetuating existing unfairness, but it's unlikely to actually create entirely new kinds of bias.
One service run in one place for 4k users has a 3 year MTBF. 1000 services run in 1000 places for 4 million users has an MTBF of about one day. How many services are the big cloud providers effectively running? They really are pretty damn reliable. It's just that a failure wipes out everyone at once, rather than each person individually having a failure that hardly anyone else ever knows about.
Obviously that doesn't mean you can just throw stuff in a cloud and forget about it. You still need your own backups and your own plans about what to do when a failure does inevitably happen. But on average, cloud services really are more reliable than on-premises in a great many cases.
"a test of the MANRS body, which was set up to fix routing errors and yet risks being dragged into global politics"
It's an international collaboration. It doesn't risk being dragged into global politics, it's deeply involved in global politics by its very nature. The technical people doing the actual work might just be trying to do their jobs properly, but you don't get a bunch of different countries arguing about how best to handle international standards and pretend that somehow politics is magically not involved in any way.
"Meanwhile in the UK... Watchdog Ofcom could get the power to fine internet giants up to 10 per cent of their annual global turnover or £18m, whichever is greater, if they do not remove illegal content or tackle legal-but-harmful material."
Yeah, I wonder where and which government it could possibly be referring to.
It tends to be easier to spot things when looking in from the outside than from the inside looking out. When it comes to planets, it's the difference between looking at a star and seeing a small wobble, and looking at the entire sphere of empty space surrounding us and hoping to see something tiny, dark, and barely moving. If you're looking at a forest from a nearby hillside, it's relatively easy to spot an extra tree growing a bit of a distance from most of the rest. If you're in the middle of the forest, it's pretty tricky to figure out where all the trees are.
"reduced the number of full reviews required per applicant by 71 percent and, by a conservative estimate, cut the total time spent reviewing files by at least 74 percent.”
Assuming courses are significantly oversubscribed, simply sending a randomly selected 71% of applications to the circular filing cabinet would have exactly the same effect. Based on my interviewing experience, you're usually looking at at least 10% of applications that would likely be worth interviewing. If you have 50 applicants for a post, you can safely throw 30 of them out without even reading them and still be left with too many good picks to actually interview. No-one really likes to admit it, but the whole selection and interview process is inherently arbitrary anyway, based heavily on personal feeling and "I know it when I see it" decisions. When it comes down to it, that's the entire point - if we could quantify the process we wouldn't need to worry about having humans read applications and perform interviews at all. So while discarding applications without even looking at them sounds like it should be a bad idea, the fact is that it would have no effect whatsoever on the final outcome - considering 20/50 applications at random is identical to only having 20 applications in the first place.
The only time you might actually have to worry about missing out on that hypothetical single perfect candidate is if there might not be enough decent candidates left in the remaining applications. But that just means you didn't have that many applications in the first place, so you don't need to worry about saving time by pre-filtering them.
"a passing resemblance to the alien structure from iconic sci-fi flick 2001: A Space Odyssey"
Different size, different shape, different colour, different material, different physical properties, different location. As far as I can tell, the supposed resemblance to the monolith in 2001 comes entirely down to the fact that both are somewhat rectangular. A box of Weetabix bears as much resemblance to an alien monolith as these silly "art" projects.
"As for being obsolete, you may want to pay attention to how many versions Apple presently supports concurrently - v4 is even still in the shops while the latest one is v6.."
Apple didn't even start making watches until 2015, and the fact that a product less than two years old is still in in the shops is hardly a spectactular endorsement of support. I know people still happily using Garmin watches from 2003, while less smart watches routinely last for decades. Apple's reputation for good support is not actually based on them being any good, but rather on everyone else who makes mobile phones being utter shit.
Maybe Apple will turn out to support their watches better than their phones, and maybe it's unfair to criticise them for making their old products obsolete when they've only been making them for a few years and haven't had time to... oh, right, they've already stopped suporting all their watches made before 2017.
Exactly what I was going to say. My parents' postcode is shared by four houses. A friend who moved into a new build briefly had the only occupied property at their postcode (although convincing anyone it actually existed at all was a bit of a struggle, and you still can't find it with Google maps). What kind of idiot thinks address, gender, date of birth and holiday information are somehow not personal data? Personal information doesn't mean every single piece of data can directly identify a specific person.
"They built an experiment and were able to take measurments 10^14 time faster than simulating the physics in a computer."
I'm glad someone else brought this up, I was starting to think I was missing something. Even reading the paper didn't clear things up (although given it has a typo within the first three words I wasn't expecting anything particuarly high quality). This has absolutely nothing to do with quantum computing. It's an actual physical measurement using a pretty standard photon counting system, which they then claim is faster than doing a simulation of said system. Well yes, no shit.
"Years ago, when you had to go to a research library or write a letter for information to be posted back to you, people read and check facts."
No they didn't. Years ago, when you had to make some effort to get information, the vast majority of people simply didn't get any information. But when they happened to be provided with information somehow, they believed it or not depending entirely on how well it aligned with their existing views. Hence travelling snake oil salesmen, religion, and so on.
The problem is not that society has changed and people no longer respect facts and experts, the problem is that society, and human nature, hasn't changed at all. The only thing that has changed is that modern communication methods make it more noticeable.
"As someone with hands and fingers that touch screens delight in ignoring, does a stylus work better?"
Yes, but with the caveat that it depends. A well implemented stylus can be far more reliable and accurate than a finger, as well as allowing extra functions via buttons and such. A poorly implemented stylus can be basically useless. By all accounts the Note ones generally work very well. From experience, the ones with HP laptops are about as useful as a small twig.
"They claim he defrauded "tens of millions" from his victims, but is only on the hook to pay back less than a double digit amount?"
The claim the gang of at least 60 people defrauded tens of millions from his victims. As the ringleader, he presumably took a larger share, but it's entirely possible the $10 million fine is the amount they think he personally made.
"*Excluding power etc."
Power, maintenance, physical space to actually put it, security... quite a few fairly important things are hidden inside that "etc.". The raw cost of just the hardware may only be a few months of rental, but a small business operating out of rented office space that needs to expand their lease to fit an extra desk and PC could find the effective cost being an awful lot more. Same if it requires additional personnel for maintenance, additional security measures against theft, higher insurance premiums, and so on.
If you're big enough that you don't really notice the additional overhead from dropping in an extra PC or two, then three months is a decent enough rule of thumb for considering the cost of the hardware. But depending on your size and circumstances, the incidental costs of those etcs can potentially be many times larger.
"Apple never said 'waterproof', they said 'water resistant'. Just as a water resistant jacket will soak through after a while, but a waterproof coat won't. The distinction has been in common use in clothing and watches for decades, the clue is in the words."
There's no such thing as waterproof; everything will let water through eventually, it's only ever a question of the pressure at which that happens. There is no such distinction between a waterproof coat and a water resistant one, they're just labelled with the actual static head they can cope with. There is also no such distinction for watches - the ISO standard actually bans any use of the term "waterproof", standards-compliant watches are only ever water resistant to some defined pressure or depth.
As for what Apple said, their words aren't actually particularly relevant. It's not made clear in this article but is elsewhere, the BBC for example, that this was actually about false advertising. Apple could have said anything they liked, the issue is that their adverts actually showed phones getting wey in real-world situations that were never tested and which the warranty did not cover. Pointless wrangling about the exact meaning of "resistant" never came into it - whether they said resistant, proof, cited an IP rating, or whatever, the problem was that their advertising gave a very different impression of what their products could handle. Using the correct weasel words in the small print doesn't protect you from that, even if we were discussing a case where the weasels were technically correct.
"Quad-level cell flash is cheaper to make than triple-level cell flash and increases storage density"
Is QLC actually cheaper to make than TLC? My impression was that it's cheaper per bit stored, but it's almost certainly more expensive to make per wafer, simply because it's more complicated. So it's not that it's cheaper to make and increases storage density, but rather that it's more expensive to make but increases storage density by more than enough to compensate for that in a finished product.
"There is one major danger in moving away from hard experimental data and towards computer prediction: You may miss unexpected new structure elements if your analysis relies on comparison to known structures."
Exactly. You may also just be wrong, given that even for known structures it's still not perfect. It's a neat idea and certainly has some useful applications. But all these reports saying that it's now just as good as actual measurements and might replace them in the future are just plain nonsense. It's no different from any other theoretical science - you can make all the predictions you like based on whatever previous knowledge you have, but at some point you're always going to have to look at the real things that actually exist.
"i.e., "yes, we get you were sort of just trying to push a warm fuzzy feeling, but the actual things you said are not, in point of fact, strictly true, and that's sort of the thing we're here to deal with, so.""
Exactly. The ruling does seem a bit harsh if you only consider the general gist of things. Black cabs have various features that actually would do pretty well at limiting the spread of diseases live covid compared to a normal car, so if they'd kept to just saying that it would have been fine. But if you're going to make claims about specific numbers and it turns out those numbers aren't actually correct, that's obviously going to be a problem to an oversight body whose entire job is checking you're not saying things that aren't actually correct.
Indeed, it always seems a bit of a weird argument that because trashing the only place with an actual livable biosphere is a bad idea, it must therefore also be wrong to exploit the resources of a bunch of barren rocks. There are all kinds of issues with the idea of permanent planetary colonies, and people can endlessly argue about which are more important and which can be solved, but the idea that using up some rock on a big rock is inherently bad is just plain weird.
One of the issues my grandfather had when he was well into his mental decline was a firm belief that the people in the TV were watching him. It's a little concerning that when the same thing eventually happens to me, no-one will actually notice since the things that used to indicate a serious detachment from reality are now just a common part of everyday life. I just need to make sure I have enough cats that the imaginary ones don't stand out, and I'll be completely prepared for my gentle descent into insanity.
"The clue is probably in the phrase "NINE year contract". Technology marches on"
On top of that, councilors march on. The contract was signed in 2016. The last council elections were in 2017, along with several by-elections since then. You can't blame people for wanting to change the terms now when they may well not have been the ones who agreed to the terms in the first place.
If you read the article a little more closely, you may notice that the tender notice was published in January. As annoying as the pandemic might be right now, it is not responsible for 2/3 of adults in the UK being overweight, and nearly 1/3 being clinically obese. The problem is not that people aren't currently going outside as much as they used, it's that they were never going outside in the first place. Whether turning maps into a game will actually do anything to help that may be up for debate, but it certainly makes no sense at all to criticise an organisation for thinking beyond the current mess and working on a different problem that existed well before and will continue to exist well after.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
