Posts by doublelayer

The year of the European Union Linux desktop may finally arrive

Re: It's the final countdown.

If you say so, but I'm still seeing a combination of wishful thinking coupled with finding reasons why politicians will treat the thing you want to happen a lot more seriously than they treat most things.

You originally told me about "change being driven by politicians of all levels", but now, you've reduced this to Denmark. Even if I accept your theory, that doesn't explain why I should have confidence in German and French politicians, and I don't, although I do have some confidence in German and French government IT employees.

It also assumes that all Danish politicians view the Greenland situation in the same way you do. Some of them might not think that the US is about to invade them because that would be ridiculous and stupid. Of course, the US has demonstrated that something being ridiculous and stupid isn't necessarily a reason not to do it, but that's far from proof that they would start an invasion, which would theoretically require all NATO members to declare war on them. They also have a lot of examples of Trump's negotiation style (start with something ridiculous in the hope of getting something large but having people think it's smaller because it's not the ridiculous thing), which would suggest that the risk of military invasion is not the one they need to worry about, with economic or diplomatic coercion being more likely attempts.

And, if the US did invade Greenland, shutting off Microsoft products to Denmark wouldn't be a very important tactic. Doing it would cause chaos, but they don't need to. The US has lots of soldiers they can drop onto the parts of Greenland that they care about. They have lots of ships and aircraft they can place near Greenland to look for reinforcements. A lot of those things are already in Greenland.

And, by the way, that doesn't have to be true for it to have an effect. The US could have a Greenland Invasion Bureau in Washington right now, and as long as some Danish politicians don't think that's realistic, they can easily not share your assumption about how urgent it is to switch away from Microsoft products. This cannot be the first time when you've seen a politician ignore something that you think is a serious problem because they either don't seem to recognize it as the threat that you do or because they don't like what they'd have to do if it was so they're choosing not to look at it to avoid having to make the hard choices. You seem quite pleased that politicians have made this choice. That they seem to have done so doesn't prove that they will stick with it or that they have technical understanding of what it is.

Web apps do make switching the desktop OS easier. The places I've worked that either had adopted Linux desktops or had a mix have usually been heavier users of web apps. It really comes down to what applications they need to do their job and what each one's compatibility is, which differs too much from organization to organization to make any generalizations about how easy it is. Also, in a large enough unit, the change can be implemented in stages. For example, if an entire local government switches, the general administration might be able to switch while the police department stays on the old system longer as they likely have separate IT departments in any case, as long as the intersection of core software supports both systems. That makes a transition easier because lessons learned by one department can be used to make the implementation elsewhere easier.

Re: It's the final countdown.

I fear you may be putting too much confidence into the statements of politicians, probably because you'd like what they're currently saying to happen. We all do it from time to time, assuming that what a politician says they care about today will match their actions later on. Unfortunately, they, and many other people, are great at sounding interested before completely changing their stance.

I don't necessarily mean that they'll be full-throated Windows supporters in a month, although some might. I mean that when difficulties start to arise, for example they're asked for some funds to build some piece of software which the system needs, doesn't run well under Linux, but can be reimplemented with two years of work from a small team, they might suddenly discover that they're not motivated to obtain that funding, so they tell the IT department to figure something else out. Or they are replaced by different politicians who decide this thing sounds like a waste of time or money and tries to cancel it. Or they're replaced by someone who is totally fine with that Linis thing, whatever computery nonsense they're on about, but won't do any work to help it along. Or they're really into technology and insist on adding some AI to the system because after all, they're building it from scratch, so why stick with something old when we could be innovating LLM governance.

Just because a politician announces that they're willing to do something doesn't mean they know what it is, why they're doing it, what they need to do to do it successfully, or anything else. The IT departments involved have the capability to do a lot of this on their own, but if you think they'll get support from politicians if they need it, I think you are being dangerously overoptimistic.

I don't know why some might have downvoted you, but I've seen this argument lots of times. Maybe eventually it will be true. It hasn't been any of the times I've seen it in the past, and I doubt it will be this time. Various groups deciding to adopt Linux have been heralded as the tipping point that was going to make Microsoft take notice, whether it was Dell shipping some computers with Ubuntu, various large companies with Linux desktops, or the ever-popular Munich switching some, no all, no none, no some again machines to something open source. All of those things happened, but it hasn't changed much.

I am not going to predict that happening again, and I tend to see most predictions as ideologically motivated. That goes in both directions. Although most of the ones I see are people who would really like Linux to take over predicting that Linux will take over, I see others predicting that any attempt to migrate will fail utterly based on almost no information other than that they assume that nobody can transition to Linux, when the major limiting factor is how much you want to make that change. A motivated organization can switch to whatever they want if they are willing to be consistent in migrating things, and it's worked in the past. I doubt every announced effort will get that far, but some of them will. I don't think that's going to make everyone else do it.

How to get free software from yesteryear's IT crowd – trick code into thinking it's running on a rival PC

Re: Check the kernel times.

The changing the clock trick has been defeated in most cases by now, but not everyone who used to design limited time trials knew that, so it did once work and might work again depending on how much the writers of the mechanism thought this through.

Similarly, many limited-time trials today are implemented purely locally, by storing a file or registry key somewhere that stores that they've already been installed. If you can find it, then deleting that file and reinstalling the trial gets you another trial period. If enough people do that, the software writers can easily make the trial check something against a server before unlocking, but it hasn't proven necessary enough for everyone to do it. It's not that the problem has no solution but that some of the people involved haven't bothered to implement it.

More trouble for authors as Meta wins Llama drama AI scraping case

"The courts have decided that the "Ai" companies are using books they bought in accordance with the licence associated with the purchase, so that's basically the same thing as conforming to GPL. Isn't it?"

No. The courts have decided that the license on the book, the one that forbids companies from doing what they did to them, actually doesn't apply so long as you bought the book, and it leaves the question open whether you can do that if you stole the book for a later trial. If I comply with the GPL, I actually have to follow its terms, which is why if I don't want to follow its terms, I don't use that code.

Starlink helps eight more nations pass 50 percent IPv6 adoption

Re: "Tuvalu’s rise [..] coincided with the arrival of Elon Musk’s Starlink"

It's about addresses that are used by Tuvalu residents, not people who bought domains from Tuvalu. Technically, they probably are counting any address registered to Tuvalu, so if Tuvalu's operating some of them for their own servers marketing .tv domains, those probably count, but if you buy a .tv address and put a server on the other end, that address won't count unless you bought the IP address from Tuvalu and you didn't.

Fresh UK postcode tool points out best mobile network in your area

Re: Didn't there used to be a link for corrections?

It's still here. On the forum page, labeled "send corrections".

French city of Lyon ditching Microsoft for open source office and collab tools

Re: Open Standards and Open Source

Except we do, in fact, have tons of audio codecs and, as long as you use an open one, what actually happens in the real world is the software looks at what it is and picks the right decoder to play the file. People who work on compression can make a new one, like when Opus was created, with almost no visible effect on users except the benefits of better audio quality and smaller file sizes. If we decided that we needed to mandate that audio formats needed to be interoperable, then you could only make minor patches and we'd still be stuck with the first allowed ones, which were much less efficient than what we count on today.

Standards in software define what you can do. A radio standard defines what kind of data you get on the signal, and when you want to transmit pictures as well as audio, you need a new method. Most of the time, the benefits of standards are so obvious that people use them without being forced. I use HTTP to exchange data, not because someone told me that's the only allowed way to send API requests, but because it's easy for many things to use it. If I felt like it, I could implement a different protocol atop TCP and standardize only on that, or I could use only IP and do my own stream control, or I could eschew IP and build my own routing system. If I had a reason to think I could do those things better, then maybe I would.

There are a few cases where mandating a standard makes sense. Your radio example is a good one, but not because of user annoyance. It is a good one because the radio spectrum is a common resource which should be managed for the public benefit, not monopolized by any particular user. What I oppose in many cases is not standards, which are great, but mandated standards which makes it very hard to change anything no matter how better the new version might be. I especially oppose this when the environment makes it really easy to choose between standards. Mandating a standard on the phone network made sense because there's only one network, so it needs to work for everyone. I have a feeling that would have happened anyway because it's less work that way, but fine, mandate it. On the internet, it's easy to have as many different communication apps as you can fit on a phone, and they can have different purposes. You can have encrypted ones, ones that do file transfer, ones that are entirely decentralized, ones that are anonymous, ones with mesh networks, and if I don't like the idea of one, then I just don't use it. If they all have to be interoperable, then I either get none of that because you can't plug an anonymous mesh network app into one run by Facebook and expect them to understand one another, or I have to take all the features together even if I'd prefer not to.

Re: Open Standards and Open Source

When did "don't mandate total compatibility" turn into "mandate different implementations of everything"? Nothing says you can't have interoperability, but that mandating that everything be interoperable makes it really hard to build anything new because the new thing won't be supported by anything else yet. We have the standards we do because they're useful, not because they're mandated. If we decided that any short-range wireless transmission system had to be interoperable with WiFi, then we couldn't have Bluetooth or those custom 2.4 GHz autoconnect protocols for wireless keyboards or LoRa or the many incompatible things that are nonetheless useful for other purposes. Without that mandate, any number of people can still use WiFi, and it can be standardized so that people can trust that WiFi-enabled equipment will be able to use it.

AFRINIC election annulled after ICANN writes angry letter to African regional internet registry

The problem with option A, which is probably why they didn't do it, is that it's not easy or cheap to bring a bunch of representatives from small companies in lots of countries in for a vote. Proxy voting has a purpose, and I'd note that you could get the same result with a one representative one vote system. Specifically, some group that stands to benefit and has enough money to arrange bulk flights to Mauritius submits fake documents saying that each one of their passengers is the single accepted representative of one voter in this election. More expensive on them, but still within their capabilities. Remember that these people were submitting fake powers of attorney, which are not minor documents to forge, so they're not trying a small crime here.

Theoretically, there is no reason this has to touch a Chinese court. China doesn't control the routing systems. It would be easy for the rest of the world to see that Afrinic has withdrawn the addresses and no longer honor the previous users' routing announcements. It might not happen that way for fear that China would decide to continue using those and we'd get some fracturing of the IP system, but it could be implemented that way.

It's not a lack of demand. It's a lack of funds. ISPs in Africa are often more starved than most other countries for IPV4 space, and it means it's hard for things like local hosting to obtain the blocks they need, with consumers often relegated to horrendous CGNAT setups. But they also have low budgets, so they can't get those addresses in a bidding war. That's one reason why they have their own RIR in the first place. Sadly, if the people running the RIR are willing to sell those outside the region they're supposed to be in, it breaks down just as much.

It's a very good thing they didn't use that method this time. The people who showed up in person didn't submit their proxy vote and decide they could go to Mauritius after all. Their proxy vote was fraudulently submitted by other people. In a similar way, people who chose not to go but went to the effort of submitting a real proxy vote ended up with two votes, and since they weren't there, there was no way to verify which one was correct. This was more than a last-minute vote change and it needs to be handled more severely.

Microsoft dangles extended Windows 10 support in exchange for Reward Points

The article covered this: because you have to pay them monthly to have enough space to store your backup, and the easiest way to get that is to buy Office365, and if you do that, maybe you'll use Office365 and then you'll want to keep having it even after you replace your computer, and that would mean plenty of cash going directly to Microsoft's bank accounts. Or maybe some people will forget they have it or not remember why and keep paying for that OneDrive forever, the way I see some people with Apple devices who have higher storage subscriptions for iCloud but don't know why or for how long. No, it's not about AI training. Sometimes, the secret profit plan is less devious and more obvious than the theories.

Tesla Robotaxi videos show Elon's way behind Waymo

Re: Gosh

"I mean they produce more cars than Waymos fleet every few days."

And, if this was a question of how many cars they could put on the streets, that might be relevant. They're only operating ten. That doesn't seem to be their problem right now. When Waymo wants to scale up, they can buy more of the cars they attach stuff to and step up the manufacturing of the hardware they attach as well, and yes, that would take lots of money, but if they have a reason to make so many, they'd probably have the money to spend on doing it.

And, if the problem is with driving quality, manufacturing capacity is completely unimportant until that's fixed. For both of them.

Re: Gosh

Maybe, if we thought the information in the article indicated things were working well, there would be more positive assessment. Care to explain why those things aren't symptomatic of major problems? Incidentally, if you read the comments on articles talking about other self-driving car efforts, you'll find that there is a lot of skepticism of all manufacturers around here. When commenters who are already less pleased with the idea than the average are faced with a manufacturer doing it with less capability than those who have come before, you'll get even more negativity. You are, of course, welcome to explain how it's more advanced than we think it is, although I should warn you that if your argument goes something like "It's more advanced you idiots, why don't you just admit it", that's not going to help.

Re: Maybe not drugs

Given that this is his third, no fifth, no seventeenth, arbitrary number reference and every single one of them included a 4, a 2, and a 0, in that order, with no digits although occasionally a decimal point between them, no.

Just say no to NO FAKES Act, EFF argues

Re: Does privacy cover this?

Not all of that is related to privacy legislation, and some of it is not correct. For example, the existing laws that forbid you from using pictures is related to restrictions on what can be done with a "likeness" which is much more similar to the "property right" the EFF is complaining about than privacy, since privacy would forbid the taking of the picture in the first place rather than just some possible uses of the content of the picture later on.

And Facebook not using user posts in the EU... that ended and the regulators are fine with that. There may be a chance that NOYB will get a judge to block that, but it's not looking great so far. Maybe a better privacy law could do what GDPR is failing to do, but this is the US which has nothing other than a few patchwork laws in a couple states that have large carve-outs and are not well-enforced in the first place. If they had privacy legislation that covered this or even required similar permission as you'd need for a model release, then the EFF would have a point that existing laws covered it. Since those don't cover it, the EFF's argument has a hole in it, and I wonder if it might work better if they suggested to remove the dangerous parts of the law rather than abandon it entirely. The protection that law is written to provide is necessary and not available elsewhere.

That goes for other countries as well, where existing privacy legislation will probably not be sufficient to deal with AI training on personal data. The logic that Facebook's successfully used to get training on text allowed under the GDPR is the same logic that would be used to train on images. We can hope that a court decides that's a big enough problem that they'll block it, or we could make it official and not gamble on the result.

Does privacy cover this?

The EFF's complaints about what's in this are quite correct. Vague statements about what is covered make this law dangerous, and virtually unlimited requirements to take down maybe almost anything when maybe anyone asks you to is a perfect way to get bad consequences. Focusing on the existence of tools that could theoretically be abused is looking in the wrong direction when focusing on those abusive uses is more important, especially when there are legitimate uses for the tools. They're not wrong about what this law does badly.

However, I have to wonder if they are wrong about what is needed to accomplish this goal. Privacy legislation would be very welcome, but I'm not convinced it would solve this problem. Privacy legislation might prevent a company from getting a bunch of pictures of someone without consent and then making a fake avatar of them, but there are lots of ways of doing that which wouldn't be covered. The problem is that AI companies have acted as if most laws don't apply to them, and unfortunately, they've just gotten a judge to say that they can use data for whatever they want with the only proviso being that they had to obtain it legally. The fact that that counts as a partial win against AI companies speaks volumes, since normally, obtaining something illegally would be an obvious problem and nobody would need to question that part of it in the first place.

So let's consider how a company could legally obtain pictures of someone who doesn't consent to them being used to create a fake video of them. Several really easy things come to mind. Maybe they're an extra or actor and agreed to appear on camera, but not to having that footage used for literally anything later on. Or what if they posted video or photographs of themselves to a public social media website? In both cases, they voluntarily chose to put their likeness online, and that would remove most of the privacy protections around it. Do we really want that to be the only bar preventing someone from using that data without their consent in such an invasive way? To prevent that, the privacy law would have to specify that any subsequent use of the data had to be agreed to beforehand, and any data produced in violation of that had to be removed. Or in other words, the privacy legislation that would fix this problem actually looks quite a lot like this legislation does.

If the EFF is so confident that some other law can fix this problem, they need to do a better job of explaining how. Until then, they only have half of an argument, which makes it a lot easier for anyone motivated to dismiss and discredit them.

Iran cyberattacks against US biz more likely following air strikes

Re: Shocked

"My understanding being the US hit the nuclear sites."

Places where they make nuclear weapons are military things. That's especially true when they're run by the military as well, so there are military members there, but even if it only employs civilians, that's a target. But if you don't think that is correct organization, basically any time one country deliberately bombs another one, it's a war. The place in Qatar that Iran bombed had US soldiers in it. They are at war and have been at least since that first bomb fell. Of course, since this isn't a neatly organized declared war, they can stop being in a war unofficially by just no longer dropping any bombs on each other; you don't need a formal peace treaty to end a war if both sides decide they're through. Maybe that will happen. It won't have stopped this thing having been a war.

And even though it's definitely a war, you can still support it. Nothing says that you can't support a war. Nothing says you can't support the aggressor in a war, although since it's usually the less popular choice, there are numerous ways to describe things such that the other side is the aggressor and you could do that with the US-Iran war, and actually quite a bit more easily than with some others. You can have and defend the point of view that war was a reasonable and appropriate reaction to previous events if you feel that way. Just know that you will have to have that argument if you support this, because war is what they did.

Re: Shocked

Whatever the purpose, bombing someone counts as a war. Israel and Iran have been in a war with one another for a while, and the US and Iran are in one now. You can argue that these weren't the first attacks of the war and that it's existed for some time, or you can argue that it just started last week, but either way, what they're doing now is called war. That doesn't mean you shouldn't have a side here. You can support one of the belligerents in a war, or you can support neither of them, and your opinion can change as events do, but it won't stop it from being one. Directly dropping bombs on military members from the other side is one of the most obvious ways to determine whether it is one, and both countries have been doing that.

Re: SSDD?

There's often a degree of restraint involved, where the attackers still do things they find useful, things where they don't think they'll get caught, or little prods in reaction to something, but they deliberately don't do something against someone who might respond by escalating. That restraint is unlikely to be seen after large bombings, so I would expect that some things may be different. Of course, they could still have some restraint if they think the retaliation might be more bombings. So far, I do think there has been said restraint, I don't think the level of activity in the last few years is the most they can do, but I don't know what they will decide to do now.

Huawei chair says the future of comms is fiber-to-the-room, which China has and the rest of us don’t

Re: FTTR? Really?

Of course some people might need or want that increased speed, and they can and will pay for as much fiber as they can take. There are various places that could use that much. For example, some places that do research on massive datasets and move them from big storage servers to big processing computers and back again would quite like to have fiber between them. However, we're not discussing whether it will be used anywhere or saying that anyone who thinks they need that speed is mistaken. We're saying that most users do not and therefore will not pay for the development of the technology.

And that is true for two reasons. Most residential users do not need that much bandwidth because they're not moving massive files. Streaming video won't use that much bandwidth either, and after a certain point, even large files download so quickly that people no longer care how quickly they moved. And those researchers I talked about in the last paragraph: they're not using it either. Oh their storage and compute systems have fiber between them just like I said, but instead of installing fiber between the server room and their lab, they put both things in the server room and they access it from a computer in their lab which only has to show the interface of an application and, at most, some real-time high-res video, which can work just fine from the copper line that was already in the building.

We will never stop demanding more performance from our technology, but most things do eventually reach a level where fewer and fewer people need more. I have 32 GB of RAM in my computer, but a lot of people don't need that, and I don't often need any more than that. Some people do, so I can't say "32 GB ought to be enough for anyone", but neither should I expect the average consumer or business user to drive down the prices of 128 GB DIMMs through mass demand. Maybe some software will come along that needs it and is used by everybody. Maybe memory makers will keep finding more efficient ways to manufacture it and I'll buy it because it's about the same cost as the lower capacity so why not have the extra capacity. But maybe neither of those things will happen and I'll stick with 32 GB and my family and friends who don't run as many VMs or big dataset crunching will stick with their 8 GB and 16 GB, because not one of them has asked me for more RAM in the last decade.

Re: An eventuality..

I disagree. Your comment demonstrates why in one way, and there's another problem with it. Here's one:

"Folks were pre-wiring CAT6 10yrs ago for the impending 10GbE rush in homes"

And how many of them have 10 Gbps uplinks from their homes? Few can get it. One of the reasons is that few who can want it. I actually saw it as an option once for a residential connection where I was helping a friend get connected, although that required unspecified additional installation charges. Did I buy that for him? No. Did I buy the 1 Gbps connection with no additional charges? No. I bought the 100 Mbps connection, because he lived alone and didn't routinely upload terabytes, so he would notice the hundredfold speed increase not at all, but he would notice the higher monthly bill every month. To some extent, I expect that hardware will improve and faster service will become the default because it's no cheaper to provide anything slower, but that's far different.

There's another problem with FTTR, assuming it means what it looks like and isn't the translation mistake people have described, namely that it doesn't work the way most users do. Most users don't want to plug their hardware into every room. They may have a few things they want cabled in, and the rest of their bandwidth is consumed over WiFi. Fiber going to a box that can handle it, attached to a few meters of copper, is not going to have such a speed difference over sending fiber those extra meters, but it's especially not going to have a speed difference for all the devices where WiFi is the bottleneck. Having fiber to each room would make it easier to have good WiFi coverage in the same way that having copper going to each room would, and in most cases, with no difference in the experience from the user.

Re: Going beyond 10Gb/s requires fiber for now

"The "build it and they will come" argument is dumb."

That is what I said in my last comment, but it is still less dumb than saying that they're the first to bring fiber lines to individual units in a multi-tenant building, which has been done in new builds and retrofits in many countries for at least a decade. I don't think we're going to see massive increases in bandwidth needs from residential users, and they'd get happier users by bringing existing quality to the many buildings that don't have fast cabling yet. Of course, there are datacenters that will need fiber going everywhere, but they tend to already have fiber going everywhere.

Re: Going beyond 10Gb/s requires fiber for now

As I said in another post, there are two options:

1. You are correct, FTTR is a mistranslation, and they are lying to claim superiority because that's often done in buildings with fiber, either actually delivering the fiber to each flat and putting something fiber-capable on the end of it or bringing fiber to a local network closet near the flat and having a short copper line from there into the actual flat.

2. You are incorrect and they actually do mean bringing fiber lines into the building and connecting them to each office or room of a house, in which case they're right that we are not doing it and maybe they are, but I don't think we'll turn out to need it any time soon.

I don't know which one it is, but number 2 is the less stupid argument, because it is very similar to the arguments for 5G and 6G they have been making for some time. The theory goes that we will eventually need more bandwidth for something, and the places with the technology to provide it will be better situated than the ones that stuck with cheaper technology that becomes a bottleneck. That was true before; houses with some internal ethernet cabling are nicer than ones that don't have that and need a mesh WiFi setup to the one location. 4G data was much faster than 3G and enabled some uses that 3G wasn't very good at, and before that, 3G was capable of data traffic at a moderate, usable speed whereas 2.5G data traffic was often either too slow or too unreliable. The problem for Huawei is that the cheap and old version of the technology we have now is actually working rather well, so to justify replacing it, we will need something that uses more bandwidth.

Re: FTTR? Really?

If that's what they mean, then they're wrong to say that China has it and other countries don't. That's still a home, whether there's one or more rooms in it. What FTTR suggests is that in something like an office, the inter-office cabling is fiber as well. There are only a few situations where that would be noticeably different from older copper cabling, but it would be noticeably more expensive.

The AIpocalypse is here for websites as search referrals plunge

Re: khtml -vs- Gecko

That's how user agents work. It's a mess of annoying cludges from history. That's why pretty much everybody's starts with "Mozilla/5.0". Sometimes, I think we'd be better off if we decided to simply discontinue using the user agent altogether. It had a point, the point is broken, just hand out the document for the URI requested.

Their point, however, was that the bot did not decide to identify itself as one. The bot is masquerading as a browser, at least by user agent.

Re: Some of hese comments seem to miss the point

Perhaps you are unfamiliar with the concept of a patent, which means there is a public document out there describing how the thing works, in a public database. That's what you have to do to get a patent. I think the company concerned probably didn't want that to be easy or they would have put their patent numbers on their website like others do, but it's still out there and available to me. Eventually, I can find that patent and read it to learn what makes their thing different from others. I was hoping to avoid having to do that manual search by using a tool which is supposed to be better at searching than I am and was probably trained on a bunch of patents including this one (from 2016) since they are public documents. It wasn't, hence why I had to do that search, which took a while, but now I have the details I was looking for.

And guess where I got that patent file eventually? I tried going to a few patent offices first, because they store the databases. Their search systems aren't very good, though. That means I ended up finding the patent number in question at...drum roll please... Google patents. You think they might have trained their own AI on that? The point being that this is yet another case where AI will try to answer a question it has no ability to answer and prove to be useless.

Re: Some of hese comments seem to miss the point

When I do searches, I don't know the answer. A wrong answer is not any better than a useless answer. In the case of your pepper grinder example, I don't expect the AI to give me a correct answer tailored to the model of grinder I have, but instead what tends to work on most grinders. Twisting the top would have been my guess to open it, and then I expect there is a hole where I put in the new peppercorns. If that's all the AI summary says, it is no better than my guess. So just to check, I put the prompt "How do I open and refill an OXO pepper grinder?" into some LLMs, and that's what I got, paragraphs to tell me to open it and put the peppercorns into the hole, which is located either on the top, bottom, or side.

Meanwhile, if I ask a question about something more specific, the AI summary is more often wrong than useless. Either way, the answer is so frequently wrong that, when it is right, I can't count on that. Meaning I have to look at results to try to figure that out which is what I was going to do anyway. Maybe you're getting better results, although the people I know who trust LLM results have often found that following those instructions has given them results they don't appreciate.

Re: Some of hese comments seem to miss the point

If you're warning us that they can always make the AI answer worse, granted. However, it's already annoying for most of us because the answers are often wrong or useless. For example, I recently deliberately ran an AI search through Google and Perplexity because I wanted to know how a certain technology works, but all normal search engines sent me to pages that had no interest in the technical details. What I really needed was to go to the effort of finding the patent for the technology or some introductory material describing the mechanism and get a summary of that, but that wasn't very easy to find in search results (including the word "patent" in the search query just found a lot of pages telling me that the mechanism concerned was in fact patented).

Seems like something that search-equipped AI agents should be able to do. They can find which pages have that information and distill the information for me, right? And I expect that they may get the information wrong, but at least Perplexity cites sources, real ones that actually exist, right? So I run the query and the LLM happily tells me that the mechanism works "electromechanically". Which is true. Yay, the LLM didn't make anything up. Of course, that's hardly any more detail than saying the thing was built out of atoms of some elements mixed together, or in other words, the AI search result was even more useless than the search results I had before, but it was pretending not to be.

In other cases, the answer is simply wrong, but people who assume that, since it looks plausible, it is probably correct end up acting on it anyway. Ideally, you get the correct answer to your query, but you're probably not getting it as often as you think.

Re: Be careful what you wish for

And you've gone too far the other way. The law and the politicians who called for it, both the Canadian version and the Australian version which was almost the same, specifically said that they had to do something about the problems of linking and quoting. If they had just stopped at quoting, maybe they'd have had a point, but they called out linking from Google and large social media as a problem requiring compensation. They were surprised when the response to a law that was passed because linking is bad was that the people they spent months saying should stop linking or pay for doing so decided to stop linking.

The problem was stronger on the social media side, where people, including the social media companies themselves, had an incentive to copy the entire text of a news article and put it on that network so people stayed there instead of going to the paper's website. Sometimes, to try to make their act less obviously illegal, they'd only copy the first four paragraphs, thus removing important context without cutting it too short that anyone was motivated to leave. Obvious copyright violation though that is, I can see why someone might decide another law is needed to do something about it. It is, as you will have noticed, entirely separate and opposite to linking, which would be the good thing they could do instead, and that's what many search engines were already doing, quoting a couple, possibly cut off, sentences for context but requiring people go to the site to read the whole thing.

And yet, the laws were written so badly that the only way to interpret them was that linking was the problem. What the backers of the laws wanted was to take money from tech and give it to newspapers, and that's why they included linking, but the result is that the law effectively tells you that including large newspapers in search results is a bad thing and you shouldn't do it.

Huawei's latest notebook shows China is still generations behind in chipmaking

Re: Still generations behind

"A 2 year old machine without AI crapware is probably faster than the latest one with it, and all the better for not having any."

That is irrelevant to any discussion of power efficiency unless that AI is absent from the old machine and completely un-turnoffable from the new one. Neither of those is true; a two-year-old machine will get any builtin AI functions the new one does and both of them can be disabled equally easily. If, at some future point because they haven't yet, they make it impossible to turn off something, they're not going to make it easy to turn it off, just for people with a different processor than yours.

Since we're talking about desktops or laptops, imagine their relative performances when wiped and a suitable Linux distro installed. No AI to complain about. Then, the actual power efficiency numbers can be considered. In most cases, the machine that's two years older will be just fine, and the difference will be small enough that it's not noticeable. That might be a little different if you're running ten thousand of them, but even then, having to buy new hardware to get that minor benefit is likely to mean you accept the costs of the less efficient processing rather than the costs of ten thousand new machines.

That isn't as true when the machines you compare to are a lot older. I've known people who were running 2009-era machines with a Core 2 Duo in them which were running at an average power consumption of about 120 W. They were getting about as much performance out of them as a Raspberry Pi 4, and possibly not needing all of that. A Raspberry Pi 4 can run in 10 W. So in that case, the efficiency savings would be noticeable and would cover the cost of the new hardware. For example, at the average UK power price of £0.245 / kWH, a power saving of 110 W, and a purchase price of, let's say, £60 for the Pi and some accessories I assume you don't already have, the time to recover costs is three months.

Re: Weasel words

I'm not sure what you're getting at, since the entire point is that TSMC, Samsung, and Intel had 7 nm earlier and now have newer methods, though of course they still use 7 nm. Also, it wasn't comparing all of Asia, but China, and specifically SMIC, to arrive at the "generations behind" claim. TSMC usually has had the most advanced manufacturing for the last few years, so that they're further along than Intel is not really a surprise.

We could also debate the relevance of that, because many cases don't mind that the chips are built on 7 nm tech. Still, if you need lower for some reason, TSMC can do it and SMIC can't yet, hence they are actually behind. Which itself is also not a surprise because until recently, China was buying many of their advanced chips from TSMC, so we've only recently seen the results of throwing a lot into getting that capacity.

Former US Army Sergeant pleads guilty after amateurish attempt at selling secrets to China

Some of them probably were, but some of that was also spammers trying to do it cheaply and not having great English skills. It also really depends on what kind of phishing this is. The kind that says you won a prize, but you need to follow some steps by talking to them is the kind that benefits most from deliberate errors, because they're sending that to millions of people and can only talk to so many to get payment details from them. The kind that gets sent to you looking for you to click a link and enter a password benefits a lot from not having those errors because they benefit from a very brief level of trust. A lot of phishing can work well without those deliberate errors, and a lot of the people using it as a tool aren't good enough at their job not to make them.

UK students flock to AI to help them cheat

Re: “um…?”

The US and UK have different versions of the pint. They're both an eighth of a gallon, which doesn't help because the two countries also have different versions of the gallon. And before some UK person insists on lambasting the US for changing everything, the US pint is the older one, and the UK one is the one they made up in 1824. And before some US person lambasts the UK for that, the US decided to have two measurements of volume, one for liquids and one for solids (if anyone tries to measure gasses in pints, I'm going to pretend they're not talking) call them both a pint, and have them be completely different. Anyone outside those countries can look down on both systems.

"But LLMs as a tool to turn an argument I've both reasearched and made, and as importantly organised, into long-form text - seems, in principal, no different to using a calculator."

I'm not sure how you arrive at that conclusion. There appear to be many major differences, all of which are incredibly important to education and most of which are similarly important to any other use anywhere.

In education, teaching writing can be as much the point as teaching the information the student is writing about. History essays are testing both of those skills, and literature classes often care or at least should care more about the writing of the essay than the analysis of whatever Shakespeare was trying to say there since the students have little experience of Shakespearean language or context and are unlikely to come up with anything interesting. Even in something as simple as a lab report, the writing is important as well as the scientific information within. Producing something adequate but being unable to do it again is exactly what that education intends to prevent.

LLMs also have the nondeterministic problem. If I use a calculator, then as long as I put the right numbers in and it isn't broken, I get the right answer. If I get the wrong answer, I did it wrong. I can put correct information into an LLM and still get wrong data out. I frequently get new evidence for this. For example, I had a situation where I needed to describe something in about two sentences, but I found it tricky to get the useful information in in less than a couple paragraphs. So I decided to see what an LLM would do with the task. Instead of taking my paragraphs and distilling something out of them, or even selecting two sentences at random from them which would have been better, it invented new incorrect information and gave me that. That's a bigger problem for those who accept whatever an LLM produces without question, but it isn't comparable to a calculator if you have to do that.

Re: Erm, but how many are getting away with it?

"I have never either agreed with, or seen the sense in requiring coursework to be submitted electronically rather than written in longhand"

Mostly because it does nothing whatsoever to prevent cheating, since it's really easy to take cheated work, whether LLM-generated or plagiarized from another source, and write it on some paper. In the past, it did make it harder to check, because comparing students' work for similarity was easy to automate, but LLMs can now generate new cheated essays for everybody so that is less easy. Still, there's no benefit, as far as I'm concerned, in making people handwrite versus type, and since most people will be expected to write by typing later on, that works for me.

If, instead, we are trying to prevent cheating, we'll probably have to worry more about where students do work than how they write what they do. It's much easier to prevent cheating when they're doing the work in a constrained and observed environment, so more tests and less homework seems like the logical start, but there are many courses that wouldn't handle that well and it involves spending less in-class time on teaching. Still, the other approaches that have been suggested haven't struck me as likely fixes. I don't hold out much hope for teachers managing to phrase assignments in a way that an LLM can't answer, I don't think AI to detect text written by AI is going to work out very well, and how they turn it in has no effect at all.

LibreOffice adds voice to 'ditch Windows for Linux' campaign

Re: 430 million

"But if Linux Mint acquires a quarter of the market it will shake Microsoft to its foundations."

I agree, but let's phrase that in a different but equally correct way:

"But if Linux Mint acquires half a billion new users and doesn't lose them within the next six months, it will shake Microsoft to its foundations."

That's not going to happen. Even if you try your absolute hardest, it's not going to happen, but most involved are not trying their absolute hardest. I am not either. I'll install Linux if requested, and I'll offer it if asked for options. People aren't asking, and when I suggest it, they aren't requesting I install it. They have asked me to install Windows 11 though, and most of the time, the computer they have does support it. This is why I'm less certain about that number of unsupported computers. There is a large number of machines that can support Windows 11 but are still running Windows 10 (let's call that X). To know how many machines in active use are going to be cut off, we need to know the value of X, but it won't show up in surveys of OS usage so we can only estimate it, and the most reliable way to estimate it is to have figures about which processors and TPMs were used in machines sold in the 2017-2020 period which I don't have and I don't really want to read.

Techie traveled 4 hours to fix software that worked perfectly until a new hire used it

Re: manuals

It's often quoted as from him, but it seems to have a somewhat clear spread from non-famous initial usage in the addiction recovery area three decades after Einstein died. Unless those quotes were made up too, it also makes more sense in that context* than it does in modern usage, where there is insanity that doesn't involve mindless repetition and mindless repetition which can better be described as stupidity than insanity.

* The first quote used "insanity" to contrast with "sanity" which was being used as a synonym of sobriety or at least a lack of addiction and related negative effects on life, and the repetition concerned involved the repeated acts of someone with an addiction even though they knew it was harming them. Suddenly, it actually makes sense to me how that quote would have come to be which makes a lot more sense than IT people dealing with frustrating users. Incidentally, although I previously knew that the Einstein source was misattributed, I only read about this initial usage about five minutes ago.

Re: manuals

I agree that it's not a good definition of insanity*, but I do think we need a word for whatever that is, because I get frustrated by that far too often. People who, because they don't know what to do, try the same thing over and over even though they have no reason to believe that anything has or even could have changed so it will work where it didn't before.

If it said "try again later", fine, keep repeating it. If you've changed something else that should unblock it, okay, run it once more. If it's exactly the same, pressing the same button on the theory that it will eventually do what you want instead of that thing that button does is a bad idea, and I wish I hadn't seen people use that approach.

* I also have trouble coming up with an actual definition for insanity. Defining mental illness is hard enough, but defining that unclear line where it goes to that higher level has often proven challenging to me. Usually, I think that the core part has to be dissociation from reality where they either invent nonexistent things or fail to recognize existent ones, but that too can be hard to define. It's a good thing I'm not a psychologist.

European consumers are mostly saying 'non' to trading in their old phones

The more likely versions are:

Trade-in site offers you £5 and sells your used device for £120. Nobody buys it because it is not worth that much.

Trade-in site offers you £5 as a discount so you buy the new device from them, then tosses the device into the bin.

Trade-in site offers you £5 sight-unseen with a little footnote about determining device condition. That makes sense, because you can't tell them everything about the device you're sending in and there could be a problem that reduces the value. They get it, immediately toss it into the bin, and tell you it is actually worth £0, but are you going to complain about such a minor loss? Are you going to pay for the thing to be shipped back?*

The alternative that gets closest to the place you're describing is where you post it on a marketplace and find a buyer, and except for any fees you might pay, the money they pay is the money you receive. Great, let's do that. The problem is that you end up spending a lot longer describing, photographing, posting, and responding to questions about a device that people probably don't want to buy because a fair price is about £25 and the lemon problem applies. That is why, in my example above, I was happier to simply hand the phone to someone who would use it than to post it for sale and deal with this work for an uncertain and small reward.

* The trade-in for value that vanishes situation is one I've experienced personally, although in that case I was expecting it. A friend's ereader had been broken, and I was buying her a replacement. They had a trade-in program and offered me a small amount for the old one. It asked me to describe damage, and it had a field for "screen broken" which I selected, but they still offered me a nominal value. To not my surprise, they determined on receiving a ten-year-old device with a mostly nonfunctional screen that they were no longer going to offer me that discount, even though in this case they had all that information up front. Of course I didn't complain, because I didn't want the broken thing back and I didn't think there was anything better to be done with it. I also wouldn't be surprised that they do the same for functional devices. It's an easy way of taking competition off the market.

Australia finds age detection tech has many flaws but will work

Re: Data collection...

"you are suggesting that "We should just sit here and suck it up"........and do absolutely nothing to protect our own privacy!"

Am I now? No, I'm suggesting two things to everybody else. One is that we focus on more things than our own privacy, since the privacy of others is actually important no matter how selfish we are. One of the best ways to avoid surveillance is to be one of a crowd. That means making sure everyone can use encryption, because practice shows that lots of people will use it if they're allowed to and that protects us all. The second thing I suggest for others is to actually think about how well their personal privacy methods will stand up to attack, because they are being attacked, and modify their plans. You have instead chosen the complacency that your methods work (somewhat) and would still do what you describe under legislation such as this (no), and those who are worried about their own privacy should be thinking further and making plans to keep it.

But, if you've decided that I'm calling you a criminal for seeking privacy when I've been condemning privacy-invading legislation, the rest of this clearly isn't getting through. In that case, you're right, your system is perfect and you'll be fine, no need to bother worrying about anybody or anything else.

Re: Data collection...

The government will be happy to decide what documents count. If a birth certificate counts, then yes, go ahead and use that. If it doesn't, then take your birth certificate down to the people who approve passports and get one of those. Either way, you will end up using paperwork the government permits for the purpose. You write as if using papers to establish identity is something governments don't know how to do. Pretty much everywhere does it, and they have their list of acceptable documents for the act you're doing. If you want to do the act and you don't have the accepted documents, then you have to go get one of them in order to do the act. Of course, the ideal document acceptance list for this particular interaction, visiting a website that has no need to identify users is "none required and none accepted if the user wants to provide one", but that's not going to happen if a law like this gets passed. Australia's considered three alternatives and since identity verification is the only one that's not really easy to circumvent*, that's where they're going to land.

* One of their other options is AI age verification, which will be really annoying and risky, but depending on what images you have to put in to be verified, it probably won't be too hard to spoof it with a quick sample from someone who looks older or even an AI adult video generator. It will also be a nonstarter because it will get a lot of the answers wrong in the opposite direction. The remaining one is trying to use a credit card, which is a lot easier to steal than valid identity documents, can be used frequently, and won't even be noticed much if they're just checking validity without charging it.

Re: Data collection...

"Only if "the system" knows that the specific identity in question is fake. How would "the system" know this?"

We appear to be in a loop. Consider again, over and over if necessary, what you would do if they told you to scan your passport. You have four logical choices:

1. Scan your passport.

2. Scan someone else's passport, knowing that simply obtaining it has made you guilty of identity theft and you can go to prison.

3. Scan a fake passport and hope they don't have the technology to immediately identify this. Most likely, they can get that if they need it and they can make your fake a crime too.

4. Refuse to use the service.

Now you are correct that some people have indeed successfully committed identity theft before, and not just people with some degree of legal cover; people in your hypothetical position, I.E. criminals, have succeeded. Maybe you'll be one of them. Others have ended up enjoying some time in a room without internet access because they tried. Some of us might think that, if you actually commit it, you belong there. It's also much harder to obtain identification documents for someone than it is to obtain a burner phone or pseudonymous (happy now? I use the right word) email address, even if you are overestimating how well those work. Other than that, you have no problem with the course of action you suggest. I am worried about what happens to everyone else, which evidently doesn't concern you, and at this point, I'm not too worried about what befalls you as you clearly have a plan you like.

Re: Data collection...

And you write as though you missed the part about verifying age by looking at identification documents and denying you access if you don't. Your burner phone (incidentally, I'm guessing you're not in Australia because Australian phone law is stricter than UK or US regulation) and emails will do no good if the website you want to access with them says "photograph of your passport or you're not getting in, and by the way a photograph of someone else's passport is identity theft and a photo of a faked passport is fraud". You can take some actions, and doing so helps, but two things remain the case and should be taken into account when we consider what laws should be:

1. Anonymity mostly works when the people you're interacting with are content to allow it. Only a few things work when the rest of the system involved is determined to block you.

2. The things you do to remain anonymous are difficult or expensive and many others don't do them, either because of the difficulty or because they don't know that you can.

Chances are that what you think will protect you isn't as strong as you think it is, but even if it's perfect, we should oppose these laws for others and so that the next push doesn't break the protections we use.

DHS warns of sharp rise in Chinese-made signal jammers it calls 'tools of terrorism'

Re: "illegal key fob jammers"

That's not a jammer. Jammers transmit to prevent something else from usefully transmitting. Your tin is a passive thing to prevent a transmitted signal from getting to where you don't want it. They're completely different and your one is legal.

Re: Don't panic people

Some of those are real concerns, for example if there is no coverage in an area, that will be a problem for emergency comms in the area. Some of them are not. For example, as long as they use radios, a jammer can jam them. It doesn't matter whether those are cellular frequencies or some other frequency band because the jammer hardware is almost identical, so unless they have massively multiband radios that can hop around to whatever band is not being jammed right now, someone with the hardware can jam emergency radio transmissions if they are willing to put in the effort. Criminals jammed traditional P2P police radios as a tactic before cellular was even an option.

Microsoft testing PC-to-Cloud-PC failover for those times your machine dies or disappears

I don't think it's about data. Mining that is not easy. Mining it and getting anything useful is very hard, since most will be random business files of no import to anyone, possibly including anyone in the business. One reason I think this is that, if I've interpreted the article correctly, they're relying on using OneDrive to store user documents, which they've been pushing for some time and occasionally gets used as a sort of backup system*, and so the data they'd be syncing is already on their servers anyway.

To me, the explanation is much simpler. If you can convince someone in IT or IT management that this works well as a backup solution, that's another item in the Microsoft bill that gets paid without question forever without anyone thinking about whether they're using it. If they actually use it, then they may start using more cloud desktops which means plenty of money per month for what you could buy for about three months rental. If they don't use it, then Microsoft gets paid for merely having the ability to provision one if requested which doesn't happen. Either way, it's easy profit.