Posts by doublelayer 9408 publicly visible posts • joined 22 Feb 2018
It must at times come up with something useful, defined as something as correct as it needed to be. However, when I or others have used it, I have seen only outcomes where it messed up or where it did something that would have been relatively easy to do without it. I wonder how often it did mess up, but the thing that it was doing was unimportant and nobody cared that it was wrong.
Large amounts of parallel compute is not a new thing. Before LLMs, there was lots of other machine learning work which people will keep doing. There's also cryptocurrency mining which has fallen lower on the hype list but there are still plenty of people spending lots of money doing it. There are animation studios that use lots of GPUs who will fund manufacturing advancements. If all of those collapse as well, there are always gamers who will need something to drive 8K displays at 144 Hz (sure, they're not doing it now because the chips can't manage it, but I know some who will if they can and I'm not sure if they'll stop there even though it seems overkill to me). GPUs are popular because they're almost as versatile as CPUs were. Everyone needs fast single-threaded performance from time to time, but although not everyone needs reasonably fast parallel compute, so many different use cases can benefit from it that they're going to keep making new chips to do it.
The one area where we agree is that individuals won't be building new raw models like GPT4. Those require too much training data and training to go somewhere. They will certainly take the ones that have already been trained and keep building things on top of those, though. I'm also not convinced that an AI winter will mean that nobody is doing LLMs anymore. I can easily see some companies deciding that they'll never make their money back and they're out, but I don't think we'll get them all to stop. Somewhere, a company will decide that there are enough clients who want to fire their remaining customer service people and are willing to pay for the LLM that does it, and those people will keep spending.
"Who and what is going to continue pushing down the cost of the chips if the massive tech giants who were driving most of those sales in the first place are no longer interested?"
Moore's law, same as before. One of two things will happen to the chips involved, mostly Nvidia's products.
1. Someone else will come up with a use case for tons of parallel compute, so they'll keep buying those chips. Nvidia will continue to receive money and invest it into faster chips. Those who want to use the chips for LLMs will be able to buy them.
2. Nobody will find any other uses for those chips. The price will fall, and improvements in manufacturing will make it easier to keep making them. Those who want to build LLMs will use more than one of them.
Option 1 is a lot more likely. Even if there was an AI winter, it's not like everyone everywhere would stop developing something around them, and even if they did, progress could still be made on those tools. Whether that progress will ever get something that can be trusted is less clear.
Is your objection that some software, including some Microsoft software, has decided to install itself into the user's directory instead of Program Files, thereby allowing an unprivileged user to install it? If that's not your complaint, what is the problem you are trying to solve? If it is your complaint, you are aware that people can install binaries into directories they have access to on Linux, that removing them from sudoers doesn't stop them doing it, that you can stop them doing it if you're willing to put heavy restrictions on what they can do, and that Windows has the same ability to take that broad an action?
The reason they won't is that's not at all a clean slate. Running atop Linux means bringing in all the things Linux has had to do for decades. That has some advantages, because they don't have to write their own implementations of filesystems or the many other components that Linux has installed, but it also means they can't do any of the things that designing from the ground up generally lets you do. They would be trading the 1990s restrictions that went into the NT kernel, many of them based on older restrictions carried over from VMS, DOS, OS/2, and other operating systems for a different set of 1990s restrictions, many of them based on older restrictions from Unix and BSD. If they want to run anything natively, they'll also have to keep around quite a few of the first set as well.
Designing that way doesn't let them do any of the things that starting from scratch and building an NNT kernel might. Not that I expect them to do that either, but if you're going to go through a long and expensive migration project, doing one that lets you design anew is probably considered more worthwhile than doing one that, if successful, means everyone is at the same place they were before and it's easier for people to stop using the product.
I've heard it before and I'll hear it again, and I probably won't believe it then either. In my estimation, Windows will stay Windows while people are still buying it and running legacy applications. If enough people stop doing that that it's not worth keeping around, then it will die, and people will use something else and use VMs or compatibility layers for emulating it when necessary. I don't think that will happen soon, as I don't know what is likely to replace it in the near future. A ground-up reimplementation seems unlikely at all, and if it happens, I see no reason why it would use a Unix or Linux base.
"When one considers what companies might put up some stiff competition against RedHat / IBM, MS is probably one of the likely candidates whom may even be acceptable to many in the OSS community."
I don't know why you assume that, and I might be wrong, but I don't think MS would be accepted. It's pretty ingrained in many free software advocates that Microsoft is the enemy, has always been the enemy, and will always be the enemy. When Microsoft builds in a Linux layer, I see people crying "embrace, extend, extinguish" at it. When Google builds Linux into operating systems, then denies users the ability to do anything with it, I see people trying to convince me that Android should be counted as a Linux system and means that open source is winning. I think that Microsoft could develop many useful tools that work well with Linux, but if they tried to make their own distribution, I expect lots of people to refuse to run it on principle.
That's true of intentional abduction, but it isn't true of accidental harm. If a child wanders away and falls down some stairs, we don't blame the owner of the stairs for not having posted a guard to monitor for unaccompanied children. There is a limit to how much we need to modify public spaces, including the internet, to attempt to get safety that will not be achieved. No matter how much we do, there will be things on the internet that a young child should not see.
It is not the intent of the ruling, though it wouldn't be hard to extrapolate it into doing that anyway. However, it is exactly what the original post in this thread would get if their idea was implemented. From previous posts, I'm guessing they're one of the people who don't like how moderators removed or posted additional information around something they agree with, and they want that to be illegal, but they haven't considered that the law they're trying to gut is the main reason that any similar posts are available at all.
"now we need to get the other side. Allow all speech and keep protections, but when you decide to block certain topics, now you're a publisher and liable."
You do realize that, with something that stupid, the law would then say that everything at all would have to be posted to keep the protections. I.E. unless you keep up the terrorist beheading videos, you're liable. You would effectively prevent all public posting, including these forums, except for those places so extreme that they don't mind hosting literally anything, no matter how illegal, that someone decided to upload. That goes for the places you like as well. Maybe you're into some conspiracy theories and you're tired of those being moderated. Sorry, but if the sites that are keeping those up ban anything, be it even more extreme ones that you don't believe in or people disagreeing, they can now be sued for anything they keep up, meaning they're much less likely to decide to keep up the stuff you want to see or get away with it if they do.
Except that you could have transferred it off that disk to something else at many times, including now if you still had the disk. Then, you could either print it out or just keep transferring the file from medium to medium as you updated your backups. I have files from decades ago that are on modern hardware because it's 135 kB, so no need to clean it out when transferring backups from the small disks to the larger disks. In many cases, the availability of data later on is not due to what it's on, but how much you cared. Lots of paper has been discarded or damaged because the stuff written on it was not stuff I valued at the time, whether I wanted it later on or not.
I've operated multi-cloud systems before to take advantage of things that were better or cheaper on a different cloud provider. I didn't really like having to do that, but if it saves a lot of money, it can make sense. Of course, it is like any other development challenge. To do it, you have to actually know about the differences, meaning that you will have to read the docs and pricing pages and set up test systems on multiple providers, not just one. That takes time and effort, so unless that's worth expending, a lot of people won't do it. In a lot of cases, companies attach themselves to one cloud provider not because they have to, not because it's the best one, but out of inertia alone.
On prem isn't really an exception to this either. My multi-cloud systems sometimes include our own server room as one of the places where servers live, and I've seen companies lock themselves into one approach that includes self-hosting the hardware, for instance one where they insisted on self-hosting hardware in a location with network links that couldn't really handle the traffic they were getting from the internet. This is why I prefer to be a programmer, not an administrator.
And another part of problem two is that there are a lot of people intent on making these programs do their work. I've had several people try to use an LLM to do something. Sometimes, it is because they are lazy and don't want to do something they're supposed to. Sometimes, it's because they don't know how unreliable it is and actually think they're being helpful. In both cases, they basically just came up with a prompt, sent it to some LLM, and copied the response without any other consideration. That's how I know they did it, because those responses have often been uselessly generic if not actively incorrect. It's not just employers who want to get out of having employees by using an LLM.
"Given that it seems likely to have been a generic error fixing it one case at a time isn't going to help."
But that's the main purpose of many of the employees AI companies have hired. They have to quickly patch prompts or predefine answers every time someone comes up with another thing that breaks them. Does it print copyrighted material when you ask it? Does it show off its training when repeating words? Can an odd phrase cause it to go into gibberish mode? Does it tell people to do dangerous or lethal things? Does it start emulating a crazy person who you would run away from? Just patch over each of those holes and a lot of people will pretend it never did those things and certainly won't do it again.
Of course, paraphrasing the original sentence is often enough to make it break again, but they're not interested in making it not do the undesirable things. They're interested in having negative news stories, a user putting in the prompt that broke something, seeing something reasonable, and decide that the news story was blowing it out of proportion. That is how we can still have people post here saying that it doesn't print copyrighted content even though it has on numerous occasions.
I think you're painting a more open source friendly picture than reality. China may not protect copyright that much, but there's a ton of closed source in China. They probably won't be protected if someone pirates it, but that is a far cry from you having access to see and modify the source, and if you did, you're not guaranteed any protection if they pursue it. Chinese companies are well-known for ignoring the licenses of open source code they use, due to the same limitations in copyright law that you have been praising. Open source isn't just the freedoms, it's also how much source you can actually access. If they never give it out, then lower copyright protections don't do anything for you.
The most compelling arguments to prefer wired over wireless headphones are these, though some of them don't apply to all headphones and will be noted:
1. One fewer battery that needs charging.
2. Less power usage overall, though for many people, it's so small anyway that they don't actually care.
3. Higher latency for wireless, although there are many modern ones that have low-latency modes which reduce this significantly. This matters for some things but not as much for listening to music.
4. Longer time to connect, although in my experience, the wired ones will eventually get there when their cable has to be tilted in a certain way for them to work.
5. More complicated steps required if using them routinely with different devices.
My provider provides free calls to not as many countries, but still quite a few. Some of them are labeled land lines only for some reason, and I'm not sure if they mean it or not. Other countries are just not on the list. Many of them are not countries I call very often. For instance, Tanzania is not on the list, and I don't know anyone there, so I'm unlikely to spend very long finding the provider that does provide free calls to Tanzania. What happens when I do meet someone with a Tanzanian phone number and I want to call them. Do I start a fast search for a provider that will give me free calls, transfer my service, then call? No, I use one of the many internet-based systems, coordinate with them so they're also using it, place a call to them, and it works fine with no paperwork. That internet-based system might be hosted on my own servers.
That's part of the problem, but far from the whole story. There are many cases where those problems are solved or unnecessary. For instance, the manufacturer has already gotten the firmware for the components and a kernel linked. At that point, newer versions of Android, and especially patched versions, can run just as well on top of that. You might eventually want a kernel update, but until you need one, there is no reason why the higher levels of the stack can't be updated. And yet, many times, manufacturers choose not to build it and their building it is often a requirement for you to get it.
But sometimes, you do need a kernel update. Many times when you do, you can't just compile the latest kernel and drop it in. That is often easier when you're compiling the latest patch version of the LTS kernel in use, which hasn't introduced changes that will break the modifications that have been made. You can't know that for certain, but many efforts to extend the life of phones have successfully managed it by updating to the latest version of the kernel. They can do that only when modifications to the kernel have been released, which although required by the license, manufacturers sometimes choose to ignore. Manufacturers could do it too, but as with the last version, they sometimes just choose not to.
But you also have the villain that people always point to, those darn SoC manufacturers with their blobs. Sometimes, that's a problem. More often, it's not a problem and it's an excuse. There are times when SoC manufacturers just stop updating their blobs. At that point, you're not necessarily out of luck, but things are harder. Let's assume that manufacturers are unwilling to take on the additional effort at that point. I'm not convinced it's as hard as is claimed, but we'll let them have it. What you often find, however, is that the SoC manufacturers have continued to provide updates but the phone manufacturers aren't using them. Lots of companies have stopped updating Android even when the same chips have been used in phones running much later versions. Chances are that, if a phone was an early user of a CPU, they probably dropped support before the SoC did.
This is why I blame Google quite a bit for that. When we buy a laptop, we don't usually have to wonder about Windows or Linux dropping security updates. Microsoft did decide to drop Windows 11 support for some hardware, which they shouldn't have, but it's specific, relatively easily-checked details about specific hardware components. Everyone gets the same patches, whether they bought from a large company or a small one, whether that company is still in business or not.
With Android, there is no technical limitation involved, and determining what software updates are available requires identifying a specific model of device, and that is just to figure out what exists right now. Google could and should have designed their operating system to receive updates from a central location, not from manufacturers, and limited manufacturer-specific software to firmware components independent of the rest of the system. Patching a vulnerability in a system app doesn't require a firmware update or even a kernel patch, so it shouldn't need the manufacturer's approval. The money to write those patches is already there; someone is making them. Manufacturers should continue to build those patches for the devices they made for simple product quality standards, the kind of standards that encourages people to buy from them, but Google should have designed it so they didn't have to for anything that's not a bug in the manufacturer's code.
"Rather than waste everyone's time, better to build a working defence against these incoming emails in the first place."
People have been trying to do that since the 1990s if not before. If you can find a better way than all the alternatives, you can probably sell it to a lot of people. The problem is that, just like all the others, your solution will either pass through some phishing or dump some important emails in a location where nobody can find them, probably both. This is why people are trained to know that it might come in and tested to make sure they're aware about what to do.
Your suggestion is yet another item in the familiar category of requesting that a computer make it impossible to do something undesired. From the earliest computers, people have been asking for them to eliminate human error. Some types of human error can be limited, but none can be eliminated and some types can't be reduced much at all. By all means keep trying, but I'm going to continue having plans for if you don't achieve 100% success.
Maybe we send and receive different types of mail. I frequently have things that I want to send. Let's take an example.
I've just found an interesting paper produced by a university. This paper can be found at https://ee.engineering.someuniversity.ac.uk/ce/2018/~efermi/archives/2016/rgaaf-pg15-38.pdf
I want to tell my colleague about this because it has something relevant in it. How should I do it. I could include the URL, but evidently, that makes me bad at communication. So maybe I should give them the last search string I used to find this from a search engine. Except that search string is confusing because I was using it to replace the absent or malfunctioning search box on the university's site, so it has a site: filter and a few words that look like I just picked them out of a bag. Also, I used DuckDuckGo and they're using Google, so their results might be different. So I tell them to go to DDG, enter this search string, select result number 5, and oh right, this PDF was not what result number 5 links to. Result number 5 links to a personal page from graduate student R Feynman who worked on a paper about something else with professor Teller. I'm not interested in that, but fortunately, Feynman mentioned when linking to Teller that Teller also works on the kind of thing I'm interested in, so I click on that link and go to a page that Teller wrote. Only there do I find the link to a different paper that Teller wrote with Professor Fermi, which is why the PDF is under Professor Fermi who I hadn't heard about before. So maybe giving a search term isn't the best option.
So instead of that, maybe I should take the link I have and see if I can find a path back to the university's home page. If I'm lucky, I may be able to tell my colleague to navigate to the home page of Some University, and if they don't know that they can always google it, then find the link to the engineering departments, then the twenty more links needed to arrive at this professor's page. So that is not much better and there's a decent chance they'll take a wrong turning and end up at the wrong URL entirely.
If I have somewhere that I want someone else to go, a URL is the way to ensure they arrive there and not somewhere else. This is why, when I tell people to download something, I always make sure to give as clear a URL as possible and never give them a search string. Frequently, when I have used other methods, bad things happen. For example, they Googled something, clicked on an ad that I didn't see because I block them, and ended up in some sketchy site that's more than happy to provide them software downloads, just not the software download I told them to use.
The argument was that email is the weakest link due to technical limitations. There are a few technical problems with email that could be improved, but in many cases, they are not necessary and not used for successful phishing attacks. Email is the most often used path into an organization not because it's technically weaker, but because that's the most common way for an external party to communicate with the organization. I can't hop on a company's internal Slack or Teams groups without hacking an account (although an attacker might hack an account so people need to know that phishing can happen there as well), but I can send them email. If email were removed, then most phishing would come through whatever external communication method replaced it.
It is popular to see phishing as a technology problem. IT people like it because we are already used to trying to solve most problems with technology because we can and it doesn't require any of the less reliable methods of trying to solve it. Non-IT people like it because it means they can tell IT to fix it and stop paying attention. It sometimes makes more sense to file it there because sometimes it is attached to things that are actually IT problems such as malware or account compromise. The problem is that it's not really an IT problem any more than scam letters were a post office problem. No matter what technology-based solution you try, the improvement will be marginal. Since nothing else is being done, by all means find any technical hammer and hit that nail. However, if you expect a lot of movement, to not have to train people to be vigilant, or to not have failures, you will never get any of those things.
Because links are very frequently necessary. You can forbid them, no problem, but all that gets you is people pasting URLs into the email. People who are used to clicking links will now copy and paste them, and they still need to be trained to look skeptically at them first or a scammer can just send their phishing message with a URL in plain text.
Of course, there are lots of emails that could stop having links in them, like my financial institutions that always tell me how to find the information that's too private to include in the email by logging in manually, but not before giving me a link to the same information that could go anywhere. I would really like it if they stopped using links. They're not going to, though, so people need to learn what to do when we can't impose our good ideas on every business we deal with.
I agree with you that there's always a limit somewhere where the temporary negative emotion you cause someone is severe enough that it's not worth doing the test on that topic. However, there are people who put that limit quite low, and I think that can be a problem, because attackers do not think the same way. If someone wants to really phish you and find that informing you of a child's death is the way to get your attention, they will do it. In fact, there have been a few people who think that "child's death" is not severe enough and use, as their phish bait, faked recording of your child being tortured live for your convenience. Just because phishers do that doesn't mean the tests should. Something that is shocking but less so should still be on the table because, if someone is going to try it, then people should be aware about it.
Where that limit is depends on a lot of variables, but I am not convinced by the people who basically think that, if people felt the least bit nervous about the content of the email, it was automatically immoral to have sent it. The same is true of communication promising a reward which annoys people when the reward turns out not to exist. Sorry, but that's one of the phisher's most common lures and has been since the time of horse-based paper-scam delivery.
California, being a very large state, has a bunch of public (government-funded) universities and they're all named University of California [city name]. Their relative quality can be debated, but it is part of the same system as some of the more famous ones such as University of California Berkeley (of BSD fame) and University of California Los Angeles. Each one has its own leadership but there's a central organization that coordinates things between them. I don't know much more than that about how the system works.
The problem is not email. It's not HTML in email. In any functioning organization, there's a need for some kind of communication. Any communication can be faked. People fall for scam phone calls, scam web adverts, scam printed letters on paper, scam SMS messages, scam messages sent over E2E encrypted messaging systems, scam social media posts, and they will fall for scam anything else that can be sent in. Some parts of email could be better, although many of the worst ones have already gotten patches to try to help, not necessarily correctly.
In all of those cases, there are technical measures that will reduce the problem, but no system, no matter how thoroughly designed can use a technical measure to eliminate the problem or the need for receivers of the communication to work with skepticism. Phishing will never go away. Users will have to recognize this fact and learn how to recognize it when it comes in. Even if we eliminate email with all its historical cludges and replace it with something else, and I'm not sure what you think that something else should be, the problem will only reduce in scale, and possibly not even permanently once the scammers adapt to the change.
Nikolai wasn't the only person to not check the switch. From the article, at least two more people did the same thing, and those at least two more people covered a total of eleven machines to Nikolai's one. Also, that kind of thing is not one that a lot of people have to consider unless they've already been warned about it. I saw where this was going, but mostly because I've heard the same story in other articles. I have seen relatively few power supplies that needed manual switching, and all of those were already switched correctly, so I don't think this was common knowledge at the time.
On another topic, I've noticed that all those stories have involved devices attached to 230(+-) V power while set to 120 V. Is the damage similarly picturesque if it's set to 240 V and plugged into 120 V and the difference is mostly due to 230 V being the more common voltage from readers of this site?
I use iPhones and Macs, but I don't think Apple has the right to tell me what I can or can't do with them. They've acknowledged that with Macs, and I think they should have to with iPhones. Many people can use Macs perfectly well, using only Apple's store for installing software. The fact that they have root access and could do something else doesn't prevent them from using them productively without it. Apple doesn't have to make their phones any harder to use in order to let other app stores work. If the other app stores are tricky to use, that's their problem, and most of the users you're talking about will not be using them anyway; all the normal apps they might install will stay in the Apple store for ease of use.
Apple does not need to change the experience for the many users who do not care about any of this, and we already know that they won't be changing any more than they have to. I think your concerns about the changes to their business practices may be unfounded.
It doesn't look that easy because any string inside the quotation marks does the same thing. Maybe there are a few tokens that, if placed in there, would do something and not crash, but if so, it's not just a missing check for an empty string but a check for being present in a list of recognized tokens.
The problem with the line is that "free as in speech" software means more freedoms than "free as in beer" software. If you make software with the freedoms involved, you will have to deal with the fact that getting free as in beer copies is going to be possible, in fact somewhat easy, and you are specifically removing your ability to do something about it.
There are a few pieces of open source software that are built to be purchased. For instance, many of the projects of FUTO are open source and they ask you for payment when they start. That is fine. However, access to the source means that, without paying them, I can get my hands on the code and remove that part. They can't forbid me from doing it. I can then take my version and distribute it to others. They can't stop me from doing that. I am not going to do either of those things because I think paying them for the work they've done is perfectly fair, but you can.
That is why, when companies like this one want more money, they start stripping out the "free as in speech" stuff as well. They want it to be illegal to not pay them for the software, and the only way they can make that happen is to take away most if not all of the freedoms that open source has brought. It is no different than any other piece of you can see the source maybe software out there. What happens if I have a copy of the source for Oracle's software, which for some products is available, and I don't like what they're doing? May I modify it? No. May I use it without their approval? No. May I distribute modifications? Definitely no. Pseudo-open licenses are the same thing masquerading badly as open source, and in the case of Cockroach, that wasn't even enough.
I'm sure someone promised that, but Cloud was never going or intended to replace the IT people who configure things. You'll still need the same amount of domain configuration whether your domain controller is in a server room you have the keys to or not. To the extent that they promised lower staff, it was staff who fix broken hardware or deal with power and aircon, not those who configure services. If your company was small enough, those are the same people anyway, but some large companies managed to have fewer people working on servers by renting them and were able to take that money and spend it on bandwidth charges instead.
As with anything else, it's an alternative that could save you money but only if you understand what it is and have calculated where it helps and where it doesn't.
My experience is slightly different. There are some things that can't be done at all in Settings and can be accomplished with Control Panel. There are some things, perhaps not as many but I've hit it several times, which can't be done in Control Panel but can be accomplished in Settings. Either way, you'll only find out by searching thoroughly through both of them. Someone has to put all the stuff in one place, and I'm not that bothered which one they pick as long as all the settings go somewhere rather than some ones being hidden because some UI designer doesn't think anyone changes them.
That's probably one of the things they'll have to work out of the regulations. You should get something if you are expected to read and respond to messages outside working hours, not just if you do. The idea of refusing to send any email outside those hours, as the article says some servers have been configured to do, is the wrong plan. Instead, they should send it and deliver them to you, but you have no requirement to read them until you're back in working hours. That doesn't mean that reading an email voluntarily automatically means they have to pay you for it. Otherwise, I could read one email per hour while outside work and count it as overtime, and that's not going to work out.
The word "socialism" has meant about twenty different things to a couple hundred different countries. Saying that one socialism is equivalent to another is basically worthless. In fact, it's so vague a term that I basically try not to use it unless my paragraph includes several more sentences clarifying exactly what I'm trying to say this time. Otherwise, you get ridiculous results like implying or stating outright an equivalence between a democratic government trying an economic policy and a dictatorship trying a policy that works in the opposite way. Both policies may be unproductive, both might actually work in some sense, but if they're very different and I try to say they're the same, the argument gets confusing, nobody understands what point I'm trying to make, and some will assume that I don't know what I'm talking about. There are several terms that I have put into a similar category.
As I said, I'd rather that we not have to make a law. We only need such a law because employers have been failing to follow what you have already indicated you agree with. The point of this law is to protect employees from abuse, unfair treatment, and unfair dismissal. If there hadn't been those abuses, we could continue with the alternative that you prefer, not having any regulation about it. Since that has not worked out, the harms of miswritten regulation are considered better than living with the abuses.
Where we disagree is in this part:
"Employers and employees already have the negotiation ability to negotiate terms and reject the terms if they are not happy."
In most cases where this has been abused, there was no negotiation. It was not that the employer asked the employee on being hired to agree to work late nights whenever their manager wanted that, the employee agreed to that, and then the employee changed their mind. In most cases, nothing related to working at odd hours was negotiated at all but the employee is still expected to do it, and if they don't, they receive poor performance reviews, other limitations to their career, or dismissal. If they complain, unclear language like "reasonable additional hours when required" is used to suggest that they can't complain about the unnegotiated and unpaid overtime. That might not even be true, but since there is no explicit protection, companies appear willing to continue to use those tactics.
Yes, there will probably be some loopholes and some problems because this is the always tricky process of trying to encode common sense into law. For instance, they'll probably have to write a patch into the law to indicate that someone shouldn't stop work at 17:00:00 even when the task they're working on could be completed in a few minutes and could cause problems when uncompleted, and if they forget, there may be some arguments about that. However, this attempt to encode common sense would not be needed if people had stuck to it in the first place. When companies have been expecting people to work at all hours even though they did not negotiate to do so, did not add to the negotiations when they expected it, and did not compensate the people being asked to work more, then it is the employer that has violated the contract that was created.
As with many other posts here, I don't mind working longer hours if the stuff I'm doing is necessary and reasonable, and I will determine whether I think it is. If they have a crisis, I'll be called in, and I will help to resolve it. If they don't have a crisis and call me in anyway, then we have a problem, and it is not enough to say that I just need to get a new job. I will consider it, but since we didn't negotiate this, putting all the responsibility on me is misplacing it.
This move is necessary because some companies have acted as if they can simply state that your working hours are all hours unless they let you take them off. If they hadn't tried to do that, we wouldn't need it. Some other companies have not tried that, and it's unfortunate that they may have extra regulatory work to deal with, but if they are being reasonable, I hope that most or all of their employees will also be reasonable and leave the execution of this law to those companies that are using their own loopholes in the contracts they've written.
Some countries have. For Russia and Iran, they don't really do so anymore because the elections they hold are either ignored or preemptively disabled, but some countries, especially the US and UK, can be blamed for having done so in the past. The type, extent, and maliciousness of the attempts is more difficult to compare. It is fair to say that western countries have failed to live up to the ideals they should have, but less fair to tar them all the same way (many democratic countries had nothing to do with any of that and would or did protest actions when they were aware of them).
In fairness to their point, they're probably thinking about using a Pi as a desktop, a use case where comparisons to a phone or laptop are more viable as people aren't using either of those for embedded systems (okay, usually not). However, they have to consider that the Pi is not only used as a desktop, and the 2 GB model is the least likely to be used as one.
Even in that case, the Pi 5 is the least likely to fit the complaint. I knew people who would tell you that various Pis worked well as desktops. Back when they were saying that about units with 512 MB or 1 GB of RAM, they weren't convincing, and whenever I asked them about it, it would turn out that they set it up like a desktop, but they didn't actually use it routinely or they used it for one application. The Pi 4 improved this significantly by having larger amounts of RAM so notoriously memory-hungry GUI software on Linux would run well, but the processing and I/O weren't the fastest. Now the Pi 5 has improved both by including a significantly faster CPU and support for SSDs, not just SDs. I don't have a Pi 5, but from the statistics and my experience with the 4, I would expect that many average users could actually successfully use it as a desktop, including multitasking, without yearning to return to something more powerful.
Their first headphone socket was pretty terrible with a lot of interference to the extent that dithering was advisable to turn it from bad to mediocre. I'm not sure if they ever fixed it, and I do have many of the other versions so I could try to find out. However, after trying to use an original Pi's headphone socket for usable audio, I gave up and got used to I2S or USB audio connections. Thus, the loss of the socket in the Pi 5 doesn't concern me because I wasn't using it before.
From your comment, I'm guessing that at some point, they fixed that. However, I'm also wondering whether you can use audio over HDMI? That's always worked in my experience, so if you're using it for a media device using the HDMI interface, can you run audio over that and connect any audio systems to the other end? That should allow more channels than a simple 3.5 mm jack, and most televisions have one of those as well. It hasn't been an option for the kind of things I built, but most of those either have no screen or a small one.
"Do we really need an enormous framework such as electron for simple applications? Honest question."
No, but I can see why people use it. Building cross-platform applications is annoying. I don't really want to spend a lot of time working on UIs, and what I'd like even less is designing UIs at least three times for different operating systems, then tying each one to the backend. Chances are that, if I try, I'm going to get annoyed and stop before all the GUIs are done, or they're going to get out of sync. That has been a problem for a while and lots of frameworks for cross-platform GUI development exist.
A lot of them have some quirks, such as not looking quite right or interacting with the system in expected ways. For example, Java's standard GUI libraries have a tendency to react oddly to keyboard commands, mostly not reacting at all when they're supposed to. That's not good. I was also looking at some software that used GTK which came with an announcement in the documentation that this would possibly work with screen reader software for the blind on Linux, but wouldn't work with it under Windows. From the sound of their warning, it sounds like a lot of users don't understand how hard that is to fix and the authors don't know how frustrating that must be for a user to hear. When considering what framework to use, you have lots of little problems like that.
Using HTML seems like a realistic option. After all, every OS can render HTML. While I don't use Electron, mostly because it uses too much JavaScript and system resources for my liking, I have often built small GUI applications by spawning a local HTTP server and opening a page rather than using something else. They're lighter on resources and I don't usually stay with that method at the end, but when I'm trying to create a UI I can rely on to work well without spending too much time on it, it's one approach that I've used. While we have problems that are that annoying, frameworks to fix it will continue to proliferate and developers who want to get something working will choose them.
"adding a social media UI to an e-mail client - E2E encrypted social media operating user to user, without a central hub."
Most social media requires a central hub at some level, not working with user-to-user. I can't see what everyone here has written if every comment has to come from its user's computer. It wouldn't be on, or I wouldn't be known to it, and therefore I couldn't read it. Even finding the servers to ask for their messages would be a problem. The closest you can get is Mastodon, and there are still lots of central hubs involved. Not everything can work individual-to-individual.
"We need stableware - particularly a new OS that is stable for 20 years. Not updating more often than Premiership teams change their manager - updates that can break more than they fix."
The goal to just right code well once and basically never change it is not new, and it's no more realistic now than it ever has been. Things change more often than that or bugs get found, whether they have security vulnerabilities or just cause the program to do wrong things. It's true of everything we have and will be more or less true for everything that's similarly capable. To avoid it, you need to give up lots of functionality.
There are open source programs that do a lot of what you'd like to do, and it's not "corporate/government restrictions, abuse of patents, lack of interest in the VC community, and industry gatekeeping" that are keeping them from taking over. In many cases, it's users not wanting to do the extra effort that comes with self-hosting something and, to some extent, lack of the funding necessary to advance them quickly as competitors can. As The Register noted, there are barriers to open projects in users' desires and the software's design that have nothing to do with big tech abuses (there are many, but they're not always to blame) or government restrictions (they talk about banning encryption a lot, but they haven't actually done it).
You're problem is that you're ignoring what laws actually say. It's fine if you disagree with some or all parts of copyright law and are perfectly willing to break them yourself or to see others do it. We might disagree about whether that's ethical or advisable, but at least we'd be on the same page. An argument that "this is against the law, but I don't care about that" is easily understood.
Your argument that, because you disagree with the law therefore that's not what the law says, is not helping. Not only do we keep telling you that it is what the law says, but it can have harmful consequences. If you insist on believing that the law is what you want the law to be, you will continue to be surprised when the real laws get applied. Admittedly, those laws aren't being applied to the large AI companies because they have a lot of lawyers, but they do have dozens of lawsuits using these laws, which don't work as you describe, which will get to court eventually. What's worse, if you keep telling people that the law's not what it is, others may find themselves going "I'm fine, definitely following the laws, wait, why did the jury just say guilty?" If you don't like the laws, you have to try to change them. Pretending you already have will limit your ability to change them.
Well-corrected. Perhaps I should have written "Getting code you don't have permission to have and compiling it is copyright infringement even if I never give it away". The context was proprietary Adobe software, but as the chain of responses grew longer, that was less and less clear. If you do have permission to have it, or if it's in the public domain and thus you don't need it, then this all goes away.
Most of your critiques are unimportant. Whether the models are storing their training data because they were ineptly developed or because that was the intent, whether they are larger than a thousandth the size of their training data for good reasons or bad, these things are not important. Were they trained on stuff the trainers did not have licenses for, and do they reiterate some of that data back. Those are important. The answers are, in nearly all cases, yes and yes. Even if you can find a model, such as your image-based model which does not quote the training data, if the answer to the first question is yes, they still have a copyright problem. You do not have the automatic right to use anything you can find on the internet. Your additional arguments to try to distract from this fact do nothing to change it.
Me: Getting code and compiling it is copyright infringement even if I never give it away.
You: No it isn't. Licenses specifically govern redistribution. You can in fact do with free/open-source code whatever you want on your own computer. You may have broken licenses acquiring it in the first place, but that's nothing to do with compiling it, and all the material LLMs are trained on were scraped on the open internet.
You're correct about open source code. Unfortunately, you're wrong about the rest of it. Open source licenses specifically allow me to use the code as I like. Non-open licenses do not give me that permission, and the copyright applies to the code. If I get the code to something proprietary without permission, no matter what I do with it, I am not allowed to have it. It is already copyright infringement. Whether I compile it or not, modified or not, it's still infringement. Copyright is not only violated when you distribute a copy you don't have permission to distribute, but also when you obtain it. If you go to piracy sites but just download, never seeding a torrent, you're still infringing. You're probably going to be ignored if the lawyers go after people, but that's because there are so many people doing it and they have bigger fish to catch. They could sue you successfully, and that applies to code as well. It doesn't matter that the data was scraped from the internet. You do not have a right to do whatever you want with any data that you can get from the internet.
"So you're talking about "we know it's the same" vs "we don't know it's not the same". Except of course we do, because as I noted in the other comment, the network simply does not have the space to store more than a thousandth of its source data. So not only are the things you're asserting very different, but we also know that the thing you're asserting about DL models cannot be true for any functioning network in anything more than a fraction of possible cases."
Your example is flawed and your conclusions are even more flawed. Your example demonstrates that the model is too small to store all of the training data. However, many other models exist that are much larger compared to their training sources. Text-based models are frequently in that category. You also have misstated the benefits of compression; a model can contain more than a thousandth of the source data if it has effectively been compressed,, text is very easily compressed, and finding the patterns that make for the best compression is exactly what neural networks do well and one of the main parts of model training. A lot of the training data may no longer be in the model, as throwing out rare or unnecessary data is another core part of training these models. The part that is still there, while a small subset of the training data, is often in there with sufficient accuracy to be quoted.
Cockroach is making money off selling licenses to this, but the people outside their companies who fix their bugs and add new features to the code get nothing. If their devs have to get paid, why don't the external ones? Maybe the license should be changed to say that, if you've made changes to the codebase, Cockroach needs to pay you for the length of time your code's been in there? Just as users have benefited from a database that you don't necessarily have to pay to use, Cockroach has been benefiting from people who do work for free, some of which probably worked at those commercial users.
I don't think that change is advisable, but nor do I think this one is. They have the right to do this, but it is an intentional bait and switch both to the users and to the contributors. I also know that, when it backfires as many similar decisions have, they'll have no choice but to extend their groups. Well, they could go open source again, but people who already abandoned open source rarely go back on their decision and are rarely trusted if they do. So while there's a revenue cap, which already conflicts with the "commercial use" provisioned in a confusing and possibly contradictory way, that cap will probably disappear at a moment's notice at some point.
If some Chinese company did try to sell their services in the US without the licenses, that's a crime. The US could have all the money that company holds in the US confiscated to pay the fines. So yes, if the company concerned did it by having people pay them through cash in the mail and hid the equipment coming in, it might work. If they try to operate like a normal business where you can just go to their website and buy it out in the open, it's not going to work. At that point, the company concerned isn't going to be making very much money for their trouble. Meanwhile, sticking to the license means they could actually sell in the market, normally, with all the typical legal protections, so they're a lot more likely to do that or ignore the US altogether. There are some regulations that are just not worth the effort to break.
This is wonderful. We've finally found something you're not endlessly negative about.
I don't know what the authors of these tools managed to do to break it. I do know that their code is all JVM-based, because when it broke I disassembled it and identified that no native libraries other than the JRE were being used. What exactly they did inside that code I can't say. I probably wouldn't have decompiled it anyway, and the users decided not to have me either reimplement or continue to troubleshoot it. We tested it on a lot of Windows 10 machines, and it reliably failed the same way on every one of them. It doesn't mean it's Java's fault, as it easily could have been a dependency's fault, but if it was, the dependency concerned was running from the JRE and most likely written in Java.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
