Posts by doublelayer 11434 publicly visible posts • joined 22 Feb 2018
A member of the US's lower house isn't called a lawmaker either. They would be a "representative". Lawmaker, however, is a generic term for a person who can make laws, which does include MPs, including the one in this article whose proposed solution is for them to make a law.
I'm not sure that distinction is true. At that age, I could easily have found ways to stay up nearly every night. If it wasn't that book sitting there on the shelf telling me that I could find out what happened next if I just took it down again, it could be and at times was any of a number of other distractions. Those things were fun, the class I needed to do in the morning was boring and out of mind at the moment, so temptation was in the anti-sleep camp. I needed to learn how to predict that and accept that I was going to defer the fun thing so the thing I didn't care about went better because it was going to be important later, and that's a really important skill to learn.
The point though is that, if they're only looking for an absolute effect on sleep, it's obvious that there will be one but not obvious whether that's a problem. Let's say that we did the full analysis it would take to prove that social media's alerts mean most children find it more alluring than books, television, or other distractions. That's not what this study can or is trying to do, but another study could. Would that justify banning children from it? Or might we need a better reason than "more children are refraining from sleep to use this than are doing so for other reasons"? I don't think that reason is good enough for any global action, though parents stopping children from doing so either by using parental control software or taking the phone at some point would be logical. I'd need to see more substantial harm than that, and given the design of social media, it's likely that more substantial harm could be demonstrated.
There are a lot of problems. Some basic ones include the self-selection and the small sample size (300 families means likely 300 test subjects, but maybe a few more for families with multiple children in the target age range). There are bigger problems though.
One of the problems is that what they're studying is not a good reason to ban something. They're studying whether social media use has an effect on sleep. We know that answer, yes, for some people it does. But that's not a good reason to ban something since there's a lot of things that have that effect especially on adolescents. In my case, it was books. I'd read them late at night, and sometimes I'd be in a good one so I'd keep reading far too late, so I'd be tired the next day with related negative effects. That would not be a good reason to ban books even though I have a feeling I'm not the only person who did that. If it was a problem that was going to have a big effect on me, my parents could have taken away my books, but another important part of growing up is learning how to manage time, sleep, and delayed gratification because that problem specifically and many like it continue through adulthood and if I get interested in something and don't sleep, there's nobody to take it away from me now and the possible negative consequences are worse.
If a social media ban is justified, it has to be for better reasons. Those better reasons exist and may be severe enough to justify banning or restricting it depending on opinion. They won't find them if they're looking at short-term (six weeks is not a lot of weeks) outcomes in a couple basic areas where the answers are likely predetermined.
That was already a big part of the article, but I think Doctor Syntax has it right. In my case, I'm not too afraid of losing my job to AI, but when pushed to use it, I am afraid that I'll be expected to speed up by using it but, if it makes mistakes, I'll be blamed for them. I can't automate something without taking the risk of consequences if the AI doesn't work, but if I automate it but still check the results to prevent that, they'll be disappointed with the lack of benefits from having automated it. This is especially true of AI stuff implemented by someone else who I don't trust to have tested how well their thing works because, at least when I write it, I have seen what its deficiencies are and can make some plans around them. This probably doesn't apply to everyone since I'm coming from a very programming-based background which also allows me to customize any AI tools I develop, but I think it's likely another of the reasons why people are averse to using this untested and mostly untestable software.
When people choose to pull in things is both not something you can change easily and not what your point argues against. Nobody at NPM or Pypi tells everyone to grab all their packages when the application is built; both things include caching specifically intended not to have them do that and easily store a single version until told by the user to fetch updates. It fetches dependencies every time when users specifically create blank environments and tell it to, nor do those environments have to pull the latest versions, though as bad as that can be, ones that pin versions of everything can be worse. Package repositories are not to blame for that user decision.
Nor do they enable that. A while ago, the GMP project's website started having problems from GitHub users. They distributed their library through traditional download-a-tarball methods and some idiots using GitHub actions decided to pull a fresh copy fifty times during each build. That caused exactly the same problem. Fortunately, since they were all using the same platform, GMP could handle it by blocking access to GitHub, but otherwise, they'd have the same kind of distributed unmanageable load that package repositories have.
And what is that? The two options we currently have are government and charitable support. Those are options that are currently available to open source, and they both help, but this entire article is about times when they don't help enough. The question is whether we can get more of those so we don't need anything else or whether that isn't enough and we need to find something else. Your comment suggests there's something we haven't thought of. And it is?
"The GPL and almost all other "open source" licenses do not say that anyone can use it for any purpose, they all impose conditions under which people can use the software,"
You are misinformed:
GPL version 3, section 2 Basic Permissions: "This License explicitly affirms your unlimited permission to run the unmodified Program."
The FSF calls this Freedom Zero. The OSI implements it in several parts of their definition. You are free to decide you disagree with both of them, but two things apply:
1. Don't tell us they didn't say or intend it when they clearly did.
2. There's a reason why, if you change that, it's not considered free software or open source. Open source does not just mean that you can see the source code. It's meant far more for a long time.
This is not correct. The contract defines the terms, but it only has power because the code is copyrighted. If the code entered the public domain, then I would be allowed to do what I please with the code regardless of the contract. When stuff enters the public domain now, people can and do ignore all terms set by the previous copyright holders. To continue allowing those restrictions if copyright was abolished, you would need to create a new legal structure to allow it. Contract law is not sufficient.
Since we're talking about open source, there's nobody required to pay and no legal method to force them to. You have no method to bring people to court for this and certainly no way to win there. If they had a license allowing that, they would have been commercial software which has always been able to do that.
This article points out the problem with your death of copyright dream. Open source software is copyrighted too, but in such a way that most uses are permitted and those uses that aren't permitted you can get away with anyway because people don't protect them. This produces a lot of public benefits, but it also produces funding crises, maintainers who get very little for all the good they do, and arguments of who should be paying and how to make that happen when there's no responsibility for them to do so. Your plan to abolish copyright just extends that problem from those who voluntarily chose to take that risk to everyone who produces anything digital. Thank you for the suggestion, but something whose only result is to make the problem we're trying to solve a lot bigger isn't going to be very useful.
You act like package repositories are a new thing. Not only are they quite old as open source goes, but they don't go away when you manually maintain your dependencies. The code still has to come from somewhere when you get copies. People could easily script a retrieval even when it was a tarball in an Apache directory listing you were supposed to manually retrieve, and even manual retrievals at the scale that people download things can be quite costly. Development and distribution are quite different, but why should one be unimportant?
You can do that, and developers already know that and choose whether to do so, but it doesn't do much about many of the problems described in the article. Getting the release of patches won't do anything to fund the servers that distribute the official version everyone else is using, nor will it ease the burden on the people doing most of the maintenance. The release of the code won't help with that either because anyone who didn't already upstream their changes isn't going to help to do so, they're just going to publish their version, so any use of the changed code is extra work for the original maintainers. I don't have a good solution to the financial and resourcing problems that many open source projects face, but getting more code out of companies, whether or not you can, isn't going to fix it.
Also, unavailable modifications is not often a big problem in comparison. In my experience, people who make changes to open source software in house are generally split into two categories:
1. People who make a massive new piece of software around an original open source component. They don't want to release their code, and many of them find a way not to by how they interact with that component. We can dislike that, and in my experience, some of the people who choose to do this are already violating licenses to manage it, but even if they started releasing all their stuff, it wouldn't do anything to ease the maintenance problem. The extra free software would be nice though.
2. People who make improvements to a piece of software, adding a feature or fixing some bugs. These people often want to upstream their changes because they don't want the maintenance burden. If the original project includes the feature they built because they need it, then it will stay working with the next update. They're already releasing their code in most cases, with stupid legal teams being the most common reason why it doesn't happen.
That would be nice. There are people it wouldn't work for. If you don't have much of an income, then it's less valuable to offset it, so someone who lost a job and filled in the gap with work on this software would benefit less, but that's a smaller problem. As alluring as the option might be, I don't think it will or even could be successfully implemented in law. Any attempt is likely to have some large loopholes or restrictions and no politician would care much because they don't understand how important this is.
"This is not true, age ranges are reported and not to everyone who asks."
Just software and websites who are mandated to ask for it if they do anything that might be restricted in age (undefined). Anyone who has another reason to want that value can get it too because there's no registration system for the API, not that one would help with anything. Or in other words, anyone who asks. I concede the age range point if we assume that everything works perfectly. And it's true that the penalties aren't on the parents or children, just the people who write software, possibly including contributors to anything open source, and websites and applications that might allow some child to see something that maybe is a problem. Nothing dangerous there.
"That's all the major commercial OSes for most commercial devices and all require more PII than this law."
I may not have been sufficiently clear. If you've got a Windows device, you can get parental control software from people other than Microsoft. If you've got an Android device, you have options other than Google. Both allow you to avoid the login requirements and both have been available for a decade or three. If you have an Apple device, that's where you've overstated the requirements since those restrictions can be implemented without any account if a parent spends about five minutes trying and, on Mac OS, you still have third-party options. These are also bad comparisons because the restrictions from every one of those companies, with their additional requirements, offer specific control over what is allowed whereas age indication does not. Any one of those parental control options lets the parent decide to deny access in hours they don't like along with a dashboard of other locks they can activate. A birth date in a box does ... nothing until another law forces it to. If the parent wants to shut off access at night let alone anything more complex, they still have to use the tools they already had.
If you were near the target age, do you think it would work the same way? Not to mention that some politicians are willing to accept that method too. Just take your picture, they'll guess at the age, if it guesses young whether correctly or not you're locked out, that picture will be treated in accordance with privacy law [data will be stored for auditing requirements or something like that and we haven't had a ransomware incident since last year].
The point remains the same; actual age bans are implemented with more than the honor system. The argument for why California's approach isn't a problem is that, since it is the honor system, we don't have a reason to worry. Not only is that incorrect, but because that system is so ineffective, it's very unlikely to stay there.
"The change is this would make them parental controls part of the first run process on starting up the device for the first time so parents will be guided into setting the controls."
The change is that the user's age is reported to everyone who asks and there are penalties for anything the government decides haven't sufficiently asked and taken action on that data. I don't consider that a thing that needs to happen, nor do I think that is justified when the benefit you're claiming is just that parents will have a slightly easier time setting up parental controls. An argument which isn't even true since all they'll have an easier time doing is entering a date of birth in a box; any rules controlling what the user is or isn't allowed to do are going to be as complex to set up as they were before. Parents have had plenty of options for controlling machines, not only including things from Microsoft, Apple, and Google. You've also overstated the requirements some of those place on people doing the controlling.
No, I am not suggesting that. I am saying that code which chooses to care about following laws would need to include checks for those laws they care about. The alternative of having a system component try to do that for them is equally invasive and more disruptive. Any project needs to decide what they're going to do with laws like this. It wouldn't be difficult for everyone to decide that they're going to ignore California's law. That's quite easy for everyone who doesn't live in California, though anyone inside California has to make the hard decision of whether to break that law or stop contributing. If, however, they decide they are not willing to ignore that law, then my original comments about where I'd put the code whose existence is already an axiom apply.
No, it's not. It is a field with nothing attached to it. The dangerous stuff will be the stuff built to attach to it. If something forces you to enter something in that field, that will be a lot more damaging. If something makes it possible for a website to read that field, that will be a lot more damaging. Systemd has not done either of those things. The law would likely require people to do them, though I wouldn't be surprised if they don't enforce it too much, but that code would be elsewhere, for my examples in the distributions' installers and in web browsers respectively.
Yes, at least the countries in which it's expected to operate and the laws applying to them. If I was told to implement age verification, I would want to do as little of that as possible since I don't want to be the one at fault when someone changes the law. Much better for that to be someone else's problem, which it is if I give the age data and let someone else decide what consequences that's supposed to have. Since I care about user privacy, I still prefer the option where the user's age is the user's business and if some software wants to know it, it's that software that can go about finding it out.
And you see why we have a problem with it. For those bans to work, people need to identify themselves. It wouldn't work if sellers of alcohol just asked people "by the way, before I sell you these, are you a child?", and if one tried, they would be in violation of the law and would be punished. So with California's current law, this does nothing. If nothing is acceptable, don't bother passing this; we've already got that. If nothing is not acceptable, they're going to take another step to enforce that the entered data is accurate and we have major problems with that, problems I'm sure you will start to ignore when they're mandated but have so far acknowledged in order to defend this useless thing.
No, this isn't what we need. All your age indication versus verification thing does is to point out that it could be worse. Helpfully, that's always true, even under New York's or the UK's laws; I could always find a way for it to be even more dystopian. That someone could conceive of a way to do it worse doesn't mean the less bad one is acceptable.
To prove that we need this, you'd need to demonstrate why. One reason a lot of us oppose this is that we think it won't stop here. But you're arguing on the basis that we don't know that so we can't base our opposition on that. I disagree, but let's go with it for now. The reason I think it will eventually become stronger is that self-identification of age means children are no more protected than they were before. Parents who would set this and prevent their children from unsetting it already had plenty of tools. Those who didn't use those tools won't use this one. Therefore, it doesn't achieve any outcome we don't already have. We don't need this not just because it will lead to negative consequences on its own*, and I'm quite confident it will, but also because it doesn't do anything. If I'm correct, then the legislators who started here will recognize that it doesn't do anything and replace it with something that does which I will dislike. If you are correct, then legislators will ignore the fact that it doesn't do anything and nobody is advantaged.
* For example, more privacy leaks for anyone who is honest to this system. This method stores the full date of birth. Ah, but you've predicted that and have a suggestion: report age groups (child/teen/adult), not dates. Great suggestion, except the law doesn't allow for age groups, just age, and if the age is reported down to the year, then when it changes, software with access to it, eventually including websites according to the legislation, can establish the birth date anyway. That could be counteracted, for example by not reporting someone has gotten older until a random number of days after their birthday, but it's not clear if that's legal and it would involve plenty more thought for a privacy advantage most programmers don't care about.
And like many other correct statements, it is often misused by people who assume it applies to everything, this author included, much like the cryptography fans it correctly mocks. There are a lot of people who are willing to surveil data who are either unable or unwilling to force users to disclose their keys. People who think their encryption which is computationally infeasible to break will protect them from the police are clueless. People who think that it's pointless to encrypt anything because someone can always force the key out of them are also clueless. There are a lot of people who think VPNs do things they don't, but that is not what XKCD 538 cautions against.
"In the case of the Civil Service job the employee doesn't have to contribute anything like what's required to get what they're promised later on."
Correct, hence why I increased the value for the private sector job such that the person in that job would receive the same value, at least if we believe the 28.97% number is a correct valuation (if we don't, then we can calculate a new number). They're receiving more than £162k because the pension is also worth something. If you compare that to someone in the private sector earning £162k and no pension, then of course the private sector number is lower.
"(whilst also having no guarantee of what they'll get)"
You could choose to buy an annuity which does have a guarantee, but there are often good reasons not to. It doesn't help to pretend the option doesn't exist.
"In either case you're not earning anywhere close to £209k so that's a pointless number to use."
Why? Whatever the eventual number when tax is calculated in, that is what the MoD candidate can earn. When comparing that job to one in the private sector, you compute the equivalent salary there, taking into account that the additional money to equal the pension contribution can be used in different ways if desired. Whenever comparing jobs, there are nonfinancial factors which should be taken into account, and many of these can be converted to money-equivalent to make comparisons easier.
I don't know if they ever said this was the right salary for the role. Perhaps it isn't, in which case the government won't get many good candidates. I've seen governments, and for that matter private companies, who insisted on paying less than it cost to get someone competent and therefore had a hard time hiring anyone and didn't like the person they ended up with.
I also don't understand this part: "Why should they forgo even a single pound of their earnings?" They should if they want the savings during retirement. That's why everyone who contributes to their own pension does it. Effectively, you could compute the matching salary by adding the £162k salary and the 28.97% pension contribution* and considering it a £209k job where you don't get a choice about how much you put into the pension, a choice that someone in the private sector earning £209k could either make the same way or to whatever extent they wanted to.
* Calculation requires adjustments for tax rates and whichever things relating to retirement are excluded from taxation, but I'm not in the UK and don't want to look up all the tax details to do that.
I'm not sure what you're disagreeing about. It sounds like you both think the pension is a large part of the compensation for the job. While your points about where the money goes and whether there's a saved amount anywhere are correct, it seems academic unless the UK government goes bankrupt.
One point I think you may be mistaken on, or at least misestimating, is the bonus potential of private sector jobs. Some of them do have massive bonuses on offer. Corporate executive staff and financial jobs especially have a lot of that. Many other jobs don't have anything like that, including many types of management. It's not useful to assume that someone has a potential bonus based on their base salary. You'd need to know at the very least what kind of job they're doing and likely also what company they're doing it at to have an idea, because there are some job roles or companies where the bonus potential is zero, even for people making plenty of cash as a base salary.
No, AI is still better. Rising oil prices is not bad because that's definitely not your fault, but it's also not great because what if oil prices went up again? Your investors would have a reason to worry whether bad things would happen again. AI means you're a wonderful example of improving efficiency and you're going to continue growing because you understand how to make the computer do all the work. It's a good thing, not just a good excuse, at least from the perspective of someone who doesn't understand what is going on.
Wonderful complaint you've got there. You apparently missed that what they wanted not to take effort was staying patched for security issues. Surveillance thrives when security is lacking because many security people see surveillance as an adversary it's their job to fight. Your de-Googled device may have the same benefit if the creator of your custom version builds updates quickly as many of them do. If it doesn't, then surveillance may have an easier time in your environment than something properly patched. When we're fighting against surveillance, one of the most effective things we can do is to make the effort to avoid it as low as possible because it helps others who otherwise wouldn't bother and, even if we don't care about anyone else, it gives us a bigger crowd to hide in.
If you don't want to do things with your phone, that's fine for you. It doesn't change what others want to do and the benefits of letting them do so securely. Many of the things you don't care about because you do exactly the same things with different hardware are completely possible in a privacy-preserving way on a smartphone. You've combined judgemental, wrong, and smug into a mixture that helps nobody.
Suggestions like this always confuse me. There's absolutely no chance either Amazon or Raspberry Pi would make anything like this because Amazon wouldn't make any money and Raspberry Pi does not and never has built mobile phones. There are people who make things that are kind of like this suggestion, so if anyone is going to, it would be them.
Part of the reason is that, even if that's what you want, it's so customized to one person's design that it isn't what anyone else wants. For example, a wrist phone for calls isn't in much demand because you can't hold it to your ear very well, so you're always on speaker. That's annoying in public to the caller, the callee, and those nearby. You can actually buy one if you like, though they're Android-based devices for small production runs, so they can run apps (the screen isn't big though so you can probably stop yourself from using that capability). You want someone to build a product that lacks features that most willing to buy it would want. That's not something that convinces the people who are able to do that, especially when you can obtain your goal by turning things off on something closest to what you want.
It is possible, but in that case, there'd be attempts to clarify what the data needed to be, subsets which the bank rejected, and plenty to back up that understanding. That is evidence that could have been provided. We're faced with two claims at most one of which can be true. Neither of the statements makes much sense to me so far, and the theory that the company intentionally wrote a vague thing which the FSFE nonetheless correctly and completely understood to mean something specific which the company explicitly denied here requires too many implausible assumptions for me to accept it without that evidence.
There's a large and obvious difference between those two things. But even if you can't recognize that, that doesn't work for the data that isn't publicly accessible or was but not legally so, especially when the companies that collected it deleted the copyright notices that restricted their actions or questioned whether to use a known illegal source, got permission, and built a tool specifically to use it.
Yes, the worker is still a fake, by virtue of having provided fake identity and location information. That fakeness could be a small problem, like if I say your furniture is made from solid wood but it's actually particle board; if you don't notice, then maybe it's not a problem, but there are definitely differences and they can prove important whether or not you knew about that beforehand. It could also be a big problem, like if I say your furniture is made of solid wood but it's actually a variant of particle board I just created which uses this new binding substance I make which easily ignites if it gets warm. North Korea's workers are mostly option 2, which is well-documented no matter how many articles you show up on pretending like they actually do the work and don't steal information and money.
It's less that humans are helping AI and that they're actively using AI to generate reports. The people who do this tend to fall in one of three categories:
1. They're submitting bugs in the hope of a bounty payment.
2. They're submitting bugs because they want the reputation points, maybe for a item on the CV but who knows.
3. They want to improve the project and deeply believe that LLMs are capable of doing that and that the output they got will be useful to some technical person who understands it even if they do not.
Captchas don't work as well as they once did. They now block users more often than they block committed bots. Your website's form is likely only of interest to humans and basic spambots who don't bother to do more, but if there was something behind it, I think you'd find your captcha isn't as impregnable as you claim it is.
A ternary computer is not capable of any computation that a binary computer is not. What it might be, with sufficient effort used to develop it, is more efficient, generally meaning faster, on some types of computations. That's essentially what we've been doing for the history of computers. Every instruction set is equally able to implement any program, but some make it easier to do by providing facilities to writers of compilers or assembly and some can run things more quickly.
Not all the things you're complaining about were even said by Samsung. That "the discontinuation should not be interpreted as a failure or strategic withdrawal" statement isn't from Samsung, it's from IDC who are predicting, accurately or not, the market segment. It would be weird for Samsung marketing to say that, which is why they didn't. A few other statements are similarly misattributed in your comment.
"Do we not face a serious problem if we are not useful anymore?"
That depends on a lot of details about what "not useful" means. If there are things that people want done and only humans can do them, then you're still useful if you can do some of those things. That continues down the stack; if a machine does things people want but a human is needed to keep the machine working properly by adding features or maintaining it, then that human is still useful even though the machine seems to be the one fulfilling the desire. If everything that people want can be done by machines, then do humans have to be useful? That's a fun philosophical thing to consider and debate, but if it starts leading in the direction of existential crises, turn around, because we're nowhere close to that happening right now and may never get there.
For one thing, they never claimed they were "anti AI", so they don't qualify as a hypocrite by your definition.
For another, that only works if you decide that someone must oppose every possible thing under the big and vague AI umbrella, and that's an unnecessarily broad requirement. A lot of AI is so often wrong that it's useless. There are times where it's often correct, but when it is wrong, it's very wrong, so it's useless. But there are also tasks where you could use it in some ways if you've got a plan and execute it well. I dislike a lot of the people who use AI, but not because they use AI. I dislike them because they use AI and don't correct for its failings or acknowledge the problems they're causing. If you generate good results with an AI step in the middle, that can be acceptable as long as you can continue to generate those good results.
I wonder if it's related to the trust people have in the results. I think everyone has a level of pleasure they get from praise, but there's also a lot of gates deciding whether you get that, with praise that doesn't qualify being off-putting. Being told you're clever by someone you respect is nice. Being told that by someone treating you like a child is irritating. Being told it by someone you think is unqualified to know whether you are or aren't clever is probably close to neutral.
If people run queries through LLMs and come to think of LLMs as intelligent, then they might appreciate praise from them because they associate it with intelligent responses. If they run queries and decide LLMs are crap, then it's irritating. I think the types of queries people run is one indicator of how much trust they have in LLMs. For instance, many of the first queries I ran against chatbots were ones I already knew the answer to because I wanted to test how accurate and complete the responses would be, something that wouldn't work as well if I didn't fully understand what they were talking about. The bots frequently failed, hence my negative impression.
Fair enough. You clearly put a lot of thought into it. I'm still not convinced a mission statement is any good. It sounds like all you needed, and in fact mostly what you did, was changing the team name to "floor R&D". The rest of the mission statement was basically "Do R&D well for the correct value of well at the time", and I hope that's sufficiently obvious that most management would not need it stated that you didn't aim to do the job badly. It does seem like being more specific would not have been possible for your case.
I much prefer your mission statement over the buzzword-laden versions I've seen from far too many others. However, it's still not very useful, just not painful. Your version says about as much as the team name does. You switched "development" to "design and build" and added a synonym for research. If the mission statement was supposed to have any value, I'd want a clearer statement of "best possible" [on which aspects, within what limits]. If those were difficult to specify, and in my experience they change a lot so perhaps that's true, then it's difficult to have a mission statement that says anything.
We probably have to divide terms into ones we don't like but are valid using their previous definitions and ones that actually don't mean anything people can agree on. Whether you like it or not, cadence has a lot of uses related to frequency and rhythm. It's been used that way in music and bicycling as others have pointed out, and it's also used in engineering (especially when referring to machines that have repetitive motion), foot races, and fencing. Someone referring to a "monthly cadence" may sound annoying, but that's one of the things that the word means; they didn't invent that. There is a marked difference between people who phrase themselves weirdly and people who probably aren't saying anything at all and are using big words to hide that.
I think the literal translation is "find a way to benefit from people who know how to do more than one thing by having them coordinate between people who are good at one of their areas of expertise". It's not the worst suggestion, but, in my experience, most who say it in the jargony way don't know how to implement it and some don't know what they've said so they can't try.
Those who don't feel the same may have a lot of reasons, but one of the most common ones has nothing to do about shame about the code. They're trying to sell it, and selling it doesn't work too well when anyone can grab a copy and run it without that. You would end up having to have a very difficult arms race to detect and block that, one that's difficult to win since any code you wrote to try to prevent people from running the code could be identified and removed since it would also need to be public. The extent of this problem is clear with all the GPL code which, although it's not hard to comply with the license, still does not. Most projects don't care too much that they're denied their contractual rights to a mention and access to tiny changes, so it doesn't get complained about a lot, but most commercial operations would be less happy with almost anyone being able to run their software for free.
There are some who espouse the same "all code should be published" idea based on an argument that nobody should be able to hide or restrict access to information of any kind, including a right to sell it. Since you've mentioned commercial licensing, I imagine you're not in that camp. However, it does make it harder to understand why you have the view you do.
The LLM companies aren't going to do what the FSF said because they don't want to, and even if they did want to, it would likely be impossible because there's a bunch of copyrighted material in that, so releasing training data under any license, no matter what the terms, would violate copyright law even more than LLM companies already have. So far, the FSF still maintains, correctly in my opinion, that a model claiming to be free software or open source must release training data, not just model weights.
I'd like someone to reject the settlement and take this to trial because I'd like to avoid the precedent that copyright law just doesn't apply to you if you're an AI company. Doing that would be expensive, though, and perhaps the FSF is not able to do it.
Ah, your true colors reveal themselves. You leap to Apple's defense even though there was nobody to attack, and despite complaining that you needed to counter "another Apple rant", you couldn't point to the rant. The best you had was a headline on a different article, a headline which is entirely correct, since, as you said yourself, "Therefore if you consider what you're getting with the Neo or iPhone 16e isn't premium, that's correct, as it's not supposed to be." That's what compromise means; in order to have the lower price, you must accept lower specs, and depending on what you want, those specs might be too low.
The point about anecdotal evidence should be obvious. You don't have any evidence and we have no reason to trust your statements. Among other things, you've had your new MacBook Neo for at most a week now and it's running a fresh install. We have no idea what state your Windows machine is in; evidently you get lots of errors which could be Windows's fault or evidence that you've got software problems from misconfiguration which could easily produce slowness. We can't do anything useful with that comparison and neither can you. If you want to compare like with like with a stopwatch to prove the difference, that could work, but it's probably not worth it.
And finally, the 256 GB complaint. You're correct, some people don't need more. But I did not say that everyone needs more. You did, however, say that people shouldn't: "256 GB local storage and can't cope? Sort your life out."
For someone who complains about "irrational hatred of Apple products", you've got a very spirited defense system for them going on whether you needed to or not. Let me guess, I'm going to be an irrational hater now, despite saying that I thought the Neo would be quite popular?
In your quest to defend Apple there, who is the person making the rant you're fighting against? You didn't reply to anyone in particular and the article didn't seem very insulting?
I agree with some of your comments, not with others. Your speed comparison is entirely anecdotal and allows anyone else to tell you that their non-Mac is faster than a Mac using similarly empty evidence. There are also people who do need more than 256 GB of storage inside their computer, not hanging off it or across a network connection they might not always have. Evidently, you either aren't one of them or are just making assumptions, but either way, learn that not everyone is like you.
I do agree that, for the price, especially from Apple, this seems to be a reasonable spec. I expect this to be quite popular for many people. I would generally recommend the upgrade to the 512 GB model since the storage is soldered and, in my experience, storage and batteries are the components most often upgraded before replacing the machine.
I would put limits on computers that corresponded to the access I wanted the users to have. If they're not trusted to install software, no root for their account. There are many pieces of software which make that easier, or you can do it by learning a bit. A single age field will not have that level of control. For example, if a child is quite young, I'm not going to let them install any application they want, just as long as its authors didn't check the "should be an adult for this thing" box. I'll check what they want to install and, if acceptable, install it for them. Eventually, they will get that access and be trusted to install things on their own. I don't need to watch them at all times to do that.
You can use IPV6 and NAT together if you're in a situation where you want that. I have done it. It worked. There's private ranges available, so much of them that you can have really complex networks in them if you want that or you can go basic and just have fc10::1, fc10::2, and so on.
NAT breaks a lot more than you're giving it credit for, but that doesn't mean you have to stop using it if you're not doing any of those things. Please stop pretending that the designers forced their opinion on you when all the things you need to create the network you've described work just fine. Your public/private address pair thing would be the bigger change.
