From a couple of lines earlier in the article:
"the company's 2021 research rather bluntly says"
It's only been leaked recently, but it's very clear that this report was making projections for the future at the time it was created.
2253 posts • joined 3 Nov 2011
"What's the unit smaller than a nanometer ?"
In SI units, the picometre.
"Because, correct me if I'm wrong, but I thought I had already read that we are approaching the physical limits of the Universe. Just as you can't do colder than absolute zero, you can't go smaller than the size of an atom."
You're very wrong. We've been studying things smaller than atoms for well over a century now - protons were first theorised over two centuries ago. And we've been studying the structure of the things inside atoms for getting on for 60 years. Manufacturing at that scale may be a little tricky with current technology, but it's certainly nothing like a physical limit imposed by the laws of the universe.
"We don't need to get rid of guns, we need to keep a handle on nutcases."
And yet the people most in favour of guns are the same ones most opposed to any kind of competent healthcare. Texas, for example, recently implemented several even more pro-guns laws, while cutting well over $100 million from the budget for mental health care. If anyone genuinely believed that guns aren't the problem because it's really a mental health issue, you'd see the NRA and the Republican party desperately clamouring for health care reform and funding. What you actually see is the exact opposite. It makes it painfully obvious what the real motivation is.
Actually, the term comes from "language of the Franks", whose various kingdoms covered most of western Europe at different times, and it was around long before any kind of united France or French language existed. The original lingua franca was a pidgin language used around large parts of the Mediterranean for trade between people who spoke many different languages.
So it's actually the exact opposite. Not only has French never been the lingua franca, but efforts like the one in the article all but guarantee it never could be.
You probably meant cryonics. Cryogenics is the science and technology of cold stuff (usually meaning liquid nitrogen temperatures and below), and is very common in all kinds of industrial uses as well as research. Cryonics is the practice of throwing corpses in the freezer and hoping magic happens.
"Wasn't this what forced so many web sites to have a pop up to accept their cookies, and most requiring additional work to reject them?"
No. Many sites chose to do that, but since it's just as illegal for the pop-ups to work that way as it is to not have them at all, clearly legal compliance had nothing to do with that decision. The real reason so many sites did that is because they want to trick the average person into blaming the law that would protect them, instead of blaming the exploitative behaviour of the sites they are being protected from.
"these systems are nothing more than statistical regression"
Indeed. Which makes this news read a little oddly. I was wondering how exactly they were planning to generate all this fake data, and the linked WSJ article provides the answer - "Anil Bhatt said the plan is to use algorithms and statistical models to generate approximately 1.5 to 2 petabytes of synthetic data". So they're going to use algorithms and statistical models to generate the data with which to train their algorithms and statistical models. A big old circle-jerk with no hit of reality allowed to intrude at any point. They go on to say that this is how they expect everyone to do "AI" in the future.
"The first was an erroneous clock setting, that caused too much fuel to be used to complete the mission, and the capsule was safely returned."
Have you forgotten about the other two critical bugs only discovered during the flight, either of which would have caused catastrophic failure? The screw-up with the clock was the only thing that saved the mission, since NASA has stated that they would not have been found if everything else had gone to plan.
"The second was valve failures, and the capsule never left the ground This flight has had further valve failures, and that's a little worrying"
By "worrying", you probably mean "expected". Because they haven't actually fixed the problem with the valves, they've just vaguely duck taped over it (very nearly literally) and said they'll get around to a proper fix later. Not only has the temporary fix clearly not worked very well, but personally I'd be very wary of getting in a Boeing-built craft that was flown in a different configuration from that which was actually tested. They're only one step away from calling it the Starliner MAX.
"Is this the right form, or is it what we got used to?"
The hidden question there is whether there is actually a difference. It's nice to imagine that there might be some platonic ideal desktop, but in the real world the best desktop is one that allows you to get things done with the minimal amount of fuss. Making it marginally easier for a total beginner to learn doesn't actually help much, because time spent as a total beginner is absolutely tiny and the time lost by experienced users also having to relearn things is likely significant. Moving things around to make it marginally quicker to launch a program again will likely make things slower for the majority of people by more than they'll eventually benefit.
It's the same as worrying about whether there could be a faster keyboard layout than Qwerty. Probably yes. But most people can already type faster than their brain actually provides input, so a marginal increase in theoretical maximum speed benefits very few people, while the time lost learning a new layout is much more significant.
The best form is often the one most people have got used to. It doesn't matter if there might be a theoretically better form; very often the pain of changing far outweighs any potential benefits. Sometimes a gradual evolution is possible, sometimes a new generation will do things differently from established ones, sometimes it's just never worth the fuss of trying to make a big change.
""A somehow surprising result was the following: despite filling email fields on hundreds of websites categorized as Pornography, we have not a single email leak," the researchers say, noting that previous studies of adult-oriented websites have relatively fewer third-party trackers than similarly popular general interest websites."
Why do so many people always seem to be surprised by this? Porn has always been at the forefront of both technology and privacy. They have a very strong interest in not leaking anything about their customers to anyone, and are mostly run by large corporations that are perfectly capable of running a competent IT setup. If you're worried about online security, it's always been the small, amateur players you need to worry about - you make think your local church group is morally superior to a porn company, but their website will be an absolute disaster area in comparison.
It's amazing how many comments always seem like little more than grumpy old people shouting at the kids to get off their lawn. Internet videos are nothing more than TV with a slightly lower cost of entry. There's no meaningful difference between a Twitch streamer and a talk show host or an all-day shopping channel. Most of the content is shit, a relatively small portion is potentially harmful, and a fair bit of it is actually quite good. And it generally seems that a lot of the complaints boil down to little more than not understanding why people would want to watch something that the complainer doesn't personally enjoy. Why would anyone want to watch someone else playing a video game? Why not do something sensible like watch someone else kick a ball?
There is an issue with regulation. But that's nothing to do with a lack of laws, it's simply a lack of enforcement. Streamers are already not allowed to advertise without saying up front that a video is sponsored*, it's just that no-one bothers actually prosecuting them when they break the law. So it may well be a good idea to a bit more on that front. But it's bizarre seeing people applauding China when they do the equivalent of banning all TV because one program broke the law once.
Twitch and others are not inherently harmful to anyone. Actually paying TV performers for their time is not inherently a bad thing. Being awake after 2200 is not inherently a bad thing. Better regulations, and better enforcement of existing ones, would probably be good. China is not doing that because they have no interest in actually making media safe, they just want to eliminate all media that is not under state control.
* Which is itself a bit odd, given that there's absolutely no requirement to do the same for product placement in films, for example.
"Russia isn't capable of mergers or acquisitions (in either direction) at this point"
So? As the article says, part of what makes this attack interesting is that it was able to operate for over 18 months. Events of the last couple of months aren't particularly relevant to an attack started a couple of years ago.
And they weren't targetting mergers and acquisitions, they were targetting all emails of employees that work on "corporate development, mergers and acquisitions, and large corporate transactions". There's plenty there that would be of interest for both regular criminals looking for a payoff and for governmental espionage. If criminals are happy to steal via fraudulent Swift transactions directly from banks, or from all kinds of fake invoices and similar directly from companies, clearly they're not interested solely in bitcoin. Large corporate transactions sounds like a fairly sensible target for someone just looking for money. While of course the major dealings of large companies is always going to be of interest to spies. It could just be corporate espionage, but simply looking at the main targets certainly doesn't do anything to dismiss all the other possibilities.
"and to provide gender-neutral compensation for at least the next three years"
They've been prosecuted by the government for not providing gender-neutral compensation, as required by law. As part of the settlement, they're now required to follow the law, but only for the next three years. Are they then exempt from obeying the law after that time? How is it necessary to include obeying the law as part of a settlement at all, let alone with a time limit on it? If they weren't already required to obey it, they wouldn't have needed a settlement in the first place.
"There are plenty of more mainstream purposeful browsers"
Are there really? Chrome, Edge, Safari. Firefox may only have a small market share compared to its peak, but it's still the 4th biggest by a long way. Depending on whose stats you look at, it may actually be 3rd overall and 2nd on desktop. You can certainly argue there may be better browsers out there, or whatever you mean by "purposeful", but Firefox is unquestionably one of the very few mainstream browsers.
It's an old idea in sci-fi. In a post-scarcity society, what else has value other than the energy needed to keep things running? Stellaris is one of the more recent examples where money is explicitly energy credits, but even Star Trek had the basic concept before someone came up with the stupid idea of re-introducing cash. Voyager in particular is fairly explicit about having an economy based on energy with the replicator ration as the basic unit of currency.
"Talk to Alexa about something, the academics found, and the auction price for related advertising opportunities goes up."
Why was any research required here? Assume for the sake of argument that Alexa works exactly as advertised - no false activations or sending your conversations to random people or anything like that. It's essentially just a speech-to-text device. Instead of going to Amazon's website and typing in "I would like to buy some ham", you just shout to your living room "Alexa, I would like to buy some ham". Now Amazon knows you are interested in buying ham. First, they return search results (vaguely) relating to ham. Then, they pass this information to their advertising system so that you can be targetted by people attempting to sell pig products. At no point is the input device used particularly relevant, other than possibly as another data point for potential targetting.
Obviously if you don't like the very concept of targetted advertising, this is a problem. But the point is that Alexa isn't doing anything different. If you are happy searching Amazon for a product, Amazon will use that search to target you with adverts. Whether you use your fingers or mouth to enter the search terms doesn't change how the process works. So why is it supposed to be a surprise, or even particularly interesting, that Amazon uses information you've given through Alexa? Of course they do, and they use it in exactly the same way as all the other information they've collected about you. These researchers haven't revealed some shady shenanigans Amazon is getting up to, they've just confirmed that yes, targetted advertising is a thing that exists.
Personally I haven't used Amazon for years because of precisely this sort of thing. But if you deliberately buy a device that exists solely to give Amazon information, and then you use it specifically in order to give Amazon information, you don't get to be surprised that they are, in fact, collecting and using that information. If you go out of your way to tell Amazon you want ham, then next time you visit them they will sell advertising space to ham-sellers. This is not new information, and has nothing to do with whether you use Alexa or not.
"TFA says "low temperature" with the implication that 30 Celsius is low."
A lot of current "biodegradable" waste actually requires an industrial reactor running at something like 80 degrees. Which means that in practice it's just more regular plastic waste because few such reactors exist. The vast majority just ends up in landfills where it will never break down because the right conditions don't occur naturally (at least away from geothermal vents and the like). 30 is low in the sense that it will be much easier to work with on an industrial scale, and therefore more likely to actually be used at all.
It's a bit like high-temperature superconductors. They don't work at high temperatures in the sense that normal humans would think of them, but they're very high, and therefore usually much easier to work with, compared to the existing alternatives.
"But why AI? What's wrong with the traditional approach of seeding a substrate with a weak suspension, incubating and selecting the colonies that grow best? Not eye-catching enough?"
You may be a few decades out of date. The traditional approach is to run simulations on a compute cluster. This isn't a new approach replacing lab work with computers, it's just switching out the traditional genetic algorithms for machine learning. What's wrong with the really old approach is that it's incredibly slow, and extremely unreliable at actually covering a significant portion of the problem space. Real world experiments are almost always one of the later steps in the process used to confirm the results of simulations these days.
Interesting that they make no attempt to deny any of the accusations. Irrelevant statements about what else they do and denial of things that were never claimed, but overall it seems to be a a very clear endorsement that everything the researchers claimed is completely correct.
"Billions of subsidies in Europe and US will hardly make a dent"
According to the figures in the article, Taiwan currently has something over 63% of the market. In three years, they will have 44% of the market. That sounds like quite a significant dent to me, especially on such a short timescale.
"Keeping stuff cold isn't a huge problem. Sure, the cryostat will be a bit complicated, but MRI scanners only lose about 4% per month. Since they typically contain 1700 litres, which is about 200 kg, and the enthalpy of vaporisation is 23.3 kJ/kg, that's 0.04 * 200 * 23.3 = 186.4 kJ/month = 70 mW."
Not really sure what you're trying to calculate here. Heat isn't lost from coolers due to loss of coolant (unless something goes really horribly wrong). Your fridge loses close to 0% of its coolant per month, but you'll notice the back still gets quite hot. That's because cooling one place can only be done by using energy to transfer heat from one place to another, and that energy is itself converted to more heat in the process. Nearly 50% of the power use of an MRI machine is the cooling system. You're around 5-6 orders of magnitude out on how much power is required for cooling.
To take an example, the LHC uses around 40MW for the cooling system that uses 120 tonnes of helium. A similar system scaled down to 200kg would then take 66 kW. That's likely a bit of an overestimate since a widely distributed system will have more losses. But even assuming an order of magnitude gain in efficiency for a small system like an MRI, you're still looking at at least kW of cooling power. If you want to build an exa-scale supercomputer, you'll be back up into the MW region again.
"Nuclear power plants are essentially uninsurable so there always needs to be government support and a public utility for it to be a goer."
Not really, it just means they need to be self-insured. That already happens in plenty of places, especially where the company needing insurance is bigger than any insurance company and so there's simply no point in getting a third party involved. Google is one of the biggest companies in the world, with hundreds of billions of cash in hand, and even more in total assets. They're significantly bigger than many governments. Nuclear power may well be off the cards for various political and legal reasons, but when it comes to the financial side Google would be better placed to handle it than almost any public utility.
"Testing for (0, 0) is one of those really obvious problematic value pairs if you are doing anything involving any numerical operations. There are so many errors associated with 0 that if you are going to test anything at all, it should be 0."
It's worth noting that this isn't simply a question of not thinking of some obvious code tests. r and s being positive integers is part of the definition of ECDSA. It's literally step 1 of the verification (from Wiki):
"1. Verify that r and s are integers in [ 1 , n − 1 ]. If not, the signature is invalid."
They didn't simply fail to parse inputs for an otherwise correct implementation, they clearly didn't bother actually reading the specification in the first place.
"The older established players will have to be very nimble (or have something very compelling which will attract alot of others, like Disney) to actually match or even beat Netflix."
No they won't, they'll just have to stop licensing their content to Netflix. This is the fundamental problem Netflix has. People signed up because they were able to see pretty much anything from any content producer without having to ever think about who made it. Now large amounts of that content is being removed and is instead available on some other service. It doesn't matter how poor that service might be, even if people don't sign up for it they're still going to leave Netflix because the content they want is no longer there.
Because others actually are the problem. Just because many people have been conditioned into accepting an abusive situation, or simply don't understand the matter at all, that does not mean everyone should just roll over and blindly accept it. Websites that insist on loading giant piles of unnecessary scripts are a real problem in terms of usability, accessibility, and security. Pointing this out is not "going out of our way to be difficult" it's just, you know, pointing out the problem.
How many training sets are in common use that haven't already been poisoned? There seems to be a new article every week or so on how easy it is to screw with ML training data, and it seems an awful lot of people use a limited number of shared public datasets because actually gathering your own data is both difficult and often illegal. What are the chances that no TLAs or criminals (but I repeat myself) have put two and two together? Even if it's difficult to poison a dataset in advance without knowing the precise result you're going to want in the future, you'd have to assume they've done it even if only as practice for future operations. Any dataset lacking a full chain of custody for every element has to be assumed to be riddled with backdoors. Even if they're not implemented in a way that's useful for a malicious actor, how can you trust any results when any random query could hit some hidden trigger?
Indeed. Except for the part about it not being moderation bots back then. The author of the Washington Post (not NYT) article looks younger than me, and automoderation of certain words was routine on virtually all forums and similar I was using 20+ years ago. If nothing else, the obvious swearwords were generally asterisked to avoid issues with search engines or other content filters that would be triggered by them.
For that matter, Ms. Taylor may want to look up the word "euphemism" in a dictionary some time. Replacing "naughty" words with seemingly innocuous ones is far from a new idea. The details of who considers a word naughty and who tries to circumvent restrictions may vary, but a teenager saying "nip nops" instead of "nipples" is no different from my grandmother saying "sugar" instead of "shit". Neither the internet or AI have anything to do with it.
"About chivo's argument (space pollution), well, whether France authorizes Starlink or not it won't change the fact the satellites are already up there, right ?"
Sure it will. Starlink has designed their satellite coverage based on assumptions about which locations and to how many people they'll be able to sell their service. Given that France is one of the larger and richer countries around, presumably they're hoping to do a decent amount of business there. Given that there are already questions about how profitable Starlink can be, even before any competition starts showing up, losing a significantly sized country full of customers can't be a good thing. It won't make the existing satellites suddenly drop out of the sky, but they're only going to be up there for a few years anyway and there will certainly be an impact on how viable it is to replace them, or on how much larger the constellation needs to be.
Of course, the point is largely moot because there's essentially zero chance that France will actually ban Starlink. This is a mix of local politics and bureaucrats requiring the proper amount of hoop-jumping, not a serious effort to have them thrown out the country entirely.
No. Flight is always going to be more energy intensive than ground travel. An electric drone could be less polluting than a diesel van, but it will lose to an electric van. And that's assuming they're carrying comparable loads; a drone that can only carry a single small item per trip will lose to a fully loaded diesel van that diverts to the arctic to club some baby seals halfway through its delivery run.
Ultimately, you can follow the money. Why don't we already travel everywhere by flying instead of driving? The technology has been available for decades. Regulations and safety are a factor, but could easily have been covered if there was any real incentive to do so. But despite plenty of efforts to commercialise flying cars, small helicopters for personal use, and so on, none of them have ever gone anywhere for anyone other than the very rich. Why? Because flying is far more expensive than ground travel. And that cost comes down to the fact that it takes more energy. Swapping internal combustion for electric motors doesn't change that.
Indeed, Twitter has a somewhat interesting history. For most of its existence it's made massive losses, and has never had any plan for how to actually make money other than assuming it would magically happen if enough users showed up. Then Trump came along and used Twitter as effectively the official mouthpiece for the US government. For two years, Twitter made a profit. Trump was shown the door, and Twitter immediately dropped to making a loss an order of magnitude larger than it used to.
It seems more successful than it really is because journalists, not just those in Silicon Valley, find it an easy source for quote from "the man in the street" without having to bother getting off their arses and actually visiting a street. The BBC has entire articles consisting of nothing more than copy and pasting crap off Twitter. But it's clear that very few normal people actually use it, and without a highly influential, controversial figure promoting it 24/7 it's clear that there's still no-one working there who has a clue how to actually get people to use it or make any money from them. Which shouldn't be surprising because it's a one-to-many platform that's potentially useful for announcements but completely pointless for any normal person to actually use for regular communication.
And that's a big problem for Musk and others who think there's a problem with bias against right wing opinions. Twitter is propped up entirely by the liberal media. Turn it into right wing echo chamber and they'll drop it pretty much overnight. No matter how big his ego might be, Musk just isn't as big a draw as Trump was; he's been posting for years with no noticeable impact on Twitter's success or popularity. Maybe if he became president he could have a similar effect, but trying to take over and change it's operations will only kill it off faster - the problem he thinks it has is not the reason it can't make a profit.
"What if someone just "defaces" a website (as proof of "lax IT practice")? What penalty/penalties apply to that?"
From the article:
"charged with one count of causing a computer to perform a function to secure unauthorised access to a program"
Simply accessing something you're not supposed to is a crime, without any need to deface anything. Given that a website displays data, defacing it to show something else would presumably also fall foul of:
"unauthorised access to a computer with intent to impair the reliability of data" and "unauthorised access to a computer with intent to hinder access to data".
White hats already face enough issues when they engage in entirely responsible investigations. The moment you actively deface anything, even if only in a relatively mild way for what you consider educational purposes, you've very definitely crossed the line. The police may put less resources into finding you if you're not trying to extort millions from large companies, but you'll still have the book thrown at you if you do get caught. And you'll deserve it.
There are few relevant points here. Firstly, no, students are not really supposed to be more clever than the average. You can get a degree in David Beckham studies, or media studies, or golf management, or all kinds of other things that don't imply or require any particular level of intelligence. Universities offer a particular type of education in particular subjects, but there's no assumption of particularly high intelligence across the majority of attendees. That's not even really a new thing. The idea of universities consisting of high-minded academics living in ivory towers was never true, in the old days it was often just a place to ship off spare children who might be needed as backup heirs and so weren't suitable for the clergy. Intelligence in students has often been the ideal, but rarely been the case in practice.
Secondly, the defining feature of students is that they do not, in fact, yet have much of an educational level. The entire point of going to university is to
drink learn. Complaining that people specifically attempting to gain an education do not yet have an education is a bit silly.
Thirdly, one word - teenagers. The majority of people in this country go to university aged 18 or 19. If you're looking for a group of people to act in a sensible, rational manner, teenaged students would be just about the worst possible choice.
Finally, it's not greed if they really do need the money. Students as a whole tend to have no savings, no income, and have been forced to go into massive debt to the tune of tens of thousands of pounds, facing the very real prospect of remaining in debt for their entire life as a result. It's hardly surprising that some of them might jump at the chance of making some easy money.
Overall, it would be difficult to find a more perfect target. Dirt poor, uneducated, immature teenagers out on their own for the first time with no idea how the real world works but with a desperate need to find money somewhere. If we weren't already familiar with the concept of students, no scammer would believe you if you tried to invent such a perfect class of victims.
I think the more important point would be that unless the hospital has enough funding to actually provide more beds, and the staff to go with them, knowing how many people are going to turn up is irrelevant. Waiting lists and ambulance response times aren't at record levels because hospitals aren't quite good enough at predicting A&E arrivals.
Biting the hand that feeds IT © 1998–2022