Posts by doublelayer

Some scientists can't stop using AI to write research papers

Re: A solution maybe?

For me, it would be much slower. I don't think there are many people who are using LLMs in this way. However, if there are people doing that, there's nothing wrong with their resulting work. I could conceive of someone who is not confident in their writing skills trying this intending to do as I've described, though I expect that even those people will often give up in disgust when trying to edit properly and go back to writing themselves.

I think those who have made the rules calling all use of AI misconduct probably feel similarly, but I also think they're not trying to enforce it at that level. Their regulation is likely aimed to make sure that someone who is exposed as having submitted a flawed paper can't get away with blaming the AI for the problems in it. I'm not convinced that they ever could get away with it, but formalizing that isn't surprising.

Re: A solution maybe?

I'm not sure it matters. The problems with AI are not things that a well-meaning user hits. If you are writing a paper, use prompts to an LLM to aid in writing, then go back and meticulously* analyze every sentence to make sure that it never says anything your research didn't show, then you're fine. The problems come when the LLM is used because the writer no longer cares about accuracy and just wants something that looks convincing. It doesn't matter that they're using an LLM instead of a ghost writer or even their own effort to write about fictional research. The only difference is that faking papers to that extent used to be hard and now it's trivial, so you'll see a lot more of it. Hacks like banning the use of AI to write papers are trying to make this point. What is misconduct now is the same kind of thing that was misconduct before.

* This comment had no LLM input, meticulously nonetheless.

UK inertia on LLMs and copyright is 'de facto endorsement'

Re: Robots.txt says hi

Because these companies have demonstrated that they definitely care. When, for example, there is a paywall in front of it, which is much stronger than robots.txt, they bypass it anyway. What makes you think that a simple rule in that file would do anything, especially when the copy they absorbed was obtained from a site where you don't control that file anyway? Some of them have said that they respect robots.txt for the general web crawls. Let's assume we believe them. It doesn't apply to extra datasets like books 1-4 or probably more at this point or the dedicated news ones. You know, the datasets where most of the copyrighted works are.

Let's see an example. From nytimes.com/robots.txt:

User-agent: Amazonbot

Disallow: /

User-agent: anthropic-ai

Disallow: /

[...]

User-agent: ChatGPT-User

Disallow: /

User-agent: ClaudeBot

Disallow: /

User-agent: Claude-Web

Disallow: /

User-agent: cohere-ai

Disallow: /

User-agent: DataForSeoBot

Disallow: /

User-agent: Diffbot

Disallow: /

User-agent: FacebookBot

Disallow: /

User-agent: Google-Extended

Disallow: /

User-agent: GPTBot

Disallow: /

[...]

So we should be pretty assured that the New York Times has no articles in the training data, right?

Lights about to go out on US Affordable Connectivity Program

Re: Solving the Wrong Problem

I don't think that's the problem. If you don't use the internet, then what communication method do you want to use instead? Voice calls? Faxes? SMS? Paper mail? All but the last one require similar infrastructure to an internet connection, and the last one is generally more expensive than anything else. I'd also point out that the internet is one of the least proprietary of those options given that nearly every part of it is an open standard, implemented with open code.

I don't think you will solve the problem of people not having internet access by trying to pretend that it's not important technology to the way we live life today. People without it will at best be rewinded to a 1980s-style experience, and if you're serious about mandating that everyone has a non-technological communication method, make that a 1940s experience. Helping people have modern technology is more likely to assist them than trying to pretend that it isn't useful.

AI Catholic 'priest' defrocked after recommending Gatorade baptism

It could be that they paid someone to plug in prompts to a couple chatbots until they got one that didn't forget its role immediately. Lots of bots essentially take an existing bot, prime it with something like "Pretend you are a priest named Justin who believes in the Catholic Church's teachings and wants others to do so", then just make sure that they start all conversations with that. A bit more prompt engineering time can add necessary statements like "Don't say you hate types or classes of people because you are trying to be convincing". I wouldn't be surprised if that's what they spent the ten thousand on and that they probably could have managed it for less.

AWS customer faces staggering charges over S3 bucket misfire

Re: Gaming the system

Given that inbound bandwidth is one of the few things they don't charge for, this doesn't seem very useful. I am charged for each put request to S3, but if I put an object there and it is a gigabyte in size, I'm not charged any more for that. I would be charged to store it, but if I simply read it and deleted it, that is very cheap. Therefore, building a protocol around free unauthorized put requests wouldn't save you much at all.

Re: This is just one example

Because we're debating about one or two servers versus cloud. A company that needs neither of those because the only employee's personal mobile phone and company email ([e-mail]) is all the communication they use isn't relevant.

Bucket names are globally unique. [name].s3.amazonaws.com. The first to register it is the one who gets to use it. There is no naming your bucket the same as mine. The first one of us to set it up gets the name. In this case, their name was evidently used as a placeholder somewhere and people didn't update it with a real address.

I don't think that's the point in the statement. This bucket demonstrates two things:

1. If you don't allow public accesses, you can still be charged for unsuccessful attempts.

2. If you do allow public accesses and an open source library uses your bucket name and people don't change it, they end up handing you gigabytes of their data. They probably don't want you to have gigabytes of their data. They have a security problem. I think this is what they were trying to say.

Open Source world's Bruce Perens emits draft Post-Open Zero Cost License

Re: define your freedom

You've misinterpreted my comment, mixing something Bruce Perens said (in quotes) with something I said. I was quoting his comments which you can read in more detail in this article. The effects of the license discussed there are, in fact, specifically applicable to companies. It is also specifically mentioned in the Open Source Definition:

The license must not restrict anyone from making use of the program in a specific field of endeavor. For example, it may not restrict the program from being used in a business, or from being used for genetic research.

I understand that you don't think corporations should have the same rights as individuals in such licenses. This means you will probably support Perens's current plan more than I do. It does not change the fact that open source licenses do not, and by the OSD are not allowed to, discriminate against companies. Post-Open licenses can and will, which is one reason I oppose them.

I can also tell that you're trying to bring corporate personhood into this. If I did, it wouldn't help your case as it is a generally used legal construct, but it is not part of my argument anyway. We can disagree about this and it sounds like we will. I just hope that you can understand what my opinion is and I yours, including the references I have made when making my arguments.

Re: Very interestingt

We could do that, and there are some examples where we have. There are some open source programs that have been widely adopted. I'm thinking of things like Firefox, Signal, Audacity, etc, where they have been designed so that the average user can use them and they have been adopted by many people in the general public. Firefox may no longer be the most used browser, but it's still used by a lot of people and doesn't take technical knowledge to run.

Getting people to build with that in mind is a challenge we should face, and it is more likely to happen if those people are effectively told to do it, which usually means paying them to do it. It doesn't work as well if people only do what they want. I know this firsthand. I've written a program which only has a CLI interface, and I know there are at least a few people who don't know how to use that, don't want to learn, and would like a GUI. But the work of writing a GUI for this is quite a bit (this program has a bunch of parameters and several of them have special parsers, so a simple form is not going to be enough). It is plenty of work for me to do and it isn't interesting, so I haven't done it yet. I know that, if I want the general public to use this, I'm going to have to get over that and actually start writing that bit.

What I don't see is how this payment model will actually encourage anyone to do that. It doesn't pay people specifically to design with the average user in mind. It doesn't exclude people who only write for technical users. Quite importantly, it doesn't help convince people to run the thing if the devs did make it easier to use, and another reason why we don't always design with that in mind is that not many average users will run it even if we did because they don't want to do the small amount of extra work that our version would necessarily have over a managed version. Consider LibreOffice. For a lot of people, this can do all the same things as Microsoft Office can. Some exceptions may exist for some power users, but for the average student, there is no important difference. I don't have Microsoft Office on my computer. I've told people about this and showed them what I have. Still, I have been asked to help them find deals on Office licenses, and when I showed them the options, they preferred paying for a non-subscription license rather than using the other version for free. This was not because LibreOffice wasn't capable of doing something; they never tried it and I know that it can do the stuff they need done. They just valued the relatively cheap price over their idea of what learning an alternative would require. If we want them to adopt an open source version, we will have to be able to explain why they should. License details won't change this.

Re: Very interestingt

I agree that we should try to fix this. I don't think your solution does, and from your statements, I'm surprised you think it does. The problems of paying developers and causing them to write user-focused software are quite independent of each other. If I'm completely wrong and the Post-Open license is a global success, it will pay developers for their work, but it won't do anything to make them write their software differently.

If you want to make open source software more popular among the general public, then you have two problems to solve:

1. Somehow get the average computer user to understand why open source, or proprietary but we try to pretend it's the same, is better than using what they already have. Your primary asset will be privacy. Your primary liability is that they will now have to pay for it (at minimum, they'll have to pay for servers where necessary and more powerful equipment where the current software would outsource to the cloud). If you can convince the general public to embrace privacy, I'm all ears. I've been banging my head against that wall for a long time, and I'm likely to keep doing it, but I can't say I've made much progress.

2. Convince developers to do the boring work making their software more user friendly. You can't count on companies doing this for them. Companies could have improved any number of applications, but few of them chose to do so. You could just ask them, but I don't think that's likely to get as much support as you'll need. My suggestion here is that you make an actual company where you pay them to do it, then try to sell the resulting software or, if you're making real open source, the hosting thereof. If you can solve problem 1, that will work fine. If you can't, it's likely to fail. I've seen a few people fail at this mission before. It hasn't stopped me from considering trying to do it from time to time, but it has stopped me from actually quitting my job to found a company competing with big tech with more expensive products.

Re: Very interesting

Depending on how you try to do it, they may be mutually exclusive, or rather they may torpedo each other. Let's use an example I mentioned: Jitsi for video conferencing. Someone replied saying that they considered and rejected it because it wasn't good enough. Fair enough; they were probably considering it for actual work purposes whereas I was setting it up for an early pandemic situation where many other things were overloaded. It looked fine to me, but I wasn't trying to run a business from it. Let's say that we relicense Jitsi under this new license. The main effect will be that nobody ever uses it again.

A business can adopt Jitsi now. If they don't like what it does, they can pay someone, either their own employee or one of the original devs, to fix that problem and keep using it. Or they could just live with it. If they have to pay 1% of revenue for this, then they have to compare it to the alternatives. Teams does not cost 1% of revenue. Google Meet doesn't cost 1% of revenue. Nothing costs that much. Your version is now extremely more expensive than anything else. The money that's supposed to make the post-open version better won't show up because the proprietary ones are both better and cheaper.

This also raises a philosophical point. If someone made this post-open videoconferencing tool, why is that any better than a proprietary tool? Open source is better in my mind because of the freedoms it provides me. Some of those freedoms are already explicitly lost with a post-open version, and I don't see how most of the remaining freedoms will last if this takes off. It no longer works as a community effort. It's just another proprietary product with more onerous license terms. There are few differences in practice between a post-open product and all the other non-open "you can give us your code for free and we can still decide not to let you run it" licenses.

Re: define your freedom

The quote was also considering corporate requirements. So have all the open source licenses. It's been clear from the start that individual use only would not qualify. Maybe you think that is wrong, in which case you are free to write a different kind of license, but there is a reason why I don't support those.

Re: Very interesting

As I understand it, the rules work as follows:

1. Company earning less than $5M, using the code but not in a product: no payment.

2. Companies under $5M: using in a product: 1%.

3. Companies above $5M, using the code anywhere: 1%.

So it sounds like the medium-sized company using it internally would have to pay around $50k per year for their internal use. This is why I don't think it will be accepted. Companies will get near that threshold and either hide the fact that they're using it or go find something else.

Re: Doesn't go far enough in some ways, goes too far in others

It looks like the original commenter understood that and thought that was a bad idea. I think that's one of the only good ideas in it, so I won't be arguing their point for them, but I don't think your clarification did much to counter their statements.

Re: Who will benefit?

Even though I oppose this license quite strongly, I don't think this is a major concern. Getting small amounts of money may complicate tax filing, but it also gives you extra money to help deal with that problem. If the amount of the payment is so low that you don't care, you are free to decline the money, not receive it as income, and therefore not have to file. How much it adds to the difficulty of calculating your taxes depends on what country you're in, but for most situations, it's one number added and one piece of paper. I don't think developers are going to have too much trouble figuring this out.

The noncommercial use only tools are a valid concern, but I wonder how many developers are relying much on these in their work. A lot of development tools are already open source, so it doesn't apply to them. If the expenses would end up being more than the revenue from doing the work, then the developer has the options to find a tool that doesn't impose that cost or to decline the money so they're noncommercial again and can use the tool. While I don't think people would do that for tax reasons, if the tool prices were more than the revenue, they might.

Re: Pay who first?

The context of that sentence means that, of the two mentioned groups: artists and stockholders, the artists get their money first. You are right that other debt would get paid before artists, but the artists payments, if any, would come before any payments to stockholders.

Re: Very interesting

The sales pitch of "I can remove your $7 million license department and all it will cost you is $3 billion per year" doesn't sound like it will convince many companies, especially because there are two reasons why their license department probably won't get any smaller at all:

1. I'm guessing that simply paying the assigned amount to the organization isn't enough and a listing of used projects and some way of figuring out how important it is to the company* will be needed to distribute funds as fairly as possible. Someone needs to track and report this. That sounds like the open source compliance department.

2. There will still be lots of software that is still open source, so the compliance department still needs to exist to deal with that.

You know this better than I do, so maybe I'm missing something.

* I assume that funds will be allocated based on how much a given project is used in the company, meaning that my 200-line library which I got someone to include in one rough internal tool won't receive as much money as the million-line project with critical security requirements used in hundreds.

Re: Very interesting

"Big firms don't like change - give them a choice of paying out some money or writing their own version of libz or openssl, they're going to pay up."

I don't agree. For example, when Elastic switched their license to a non-open one so that cloud providers couldn't use it anymore, AWS could have opened a negotiation, figured out how much to pay, and given them the money to keep hosting Elastic Search. They could have afforded that easily. If Elastic tried to bargain too aggressively, AWS could have just bought out the company. The recent sale to IBM is about the same as one year of 1% revenue collection. Amazon could have owned Elastic entirely, meaning that not only do they not have to pay any more fees, but they could start to collect them from other cloud providers instead. Amazon did not choose to do either of these things. They forked and supported it, allowing other contributors to help them, but still taking on plenty of extra work requirements. They did this because it costs less to pay some people to run this, including some of the people who contributed back when they didn't own it, than to buy or license it. If the price is high enough, companies will choose work over money, and 1% of revenue, while it looks small, really isn't.

Re: Very interesting

The fee is supposed to be 1% of revenue, so the cloud provider, unless they are also making only £4M, will be paying more. As I described above, I think that will mean that cloud providers end up making sure that projects like this are not used in their systems and their doing it will make sure they're not used in lots of other ones as well.

Re: Very interesting

So go proprietary. We already acknowledge that public good by having copyright, and it's a lot longer and more powerful than patents. You can write whatever you want and sell it to anyone you want. The one thing you shouldn't do if you want to sell something is give it away first.

I've written code as open source, but I've also written proprietary stuff. There have been projects where I thought I might be able to sell them and that I'd rather do that than let anyone use and modify it for free. Advantage: I could get paid, and not through external schemes but for the work I'd already done. Disadvantage: I wasn't going to get free work from contributors or the free publicity from lots of users picking up the free version. I was free to decide how I'd sell this. I could let the buyers see the code or keep it private. I could let buyers modify the code or block them. But it was still proprietary, because in order to make sure that I could sell it, I didn't grant anyone a license to use, modify, and distribute the way that I do when I put an open source license on things. If they wanted to use it, they had to pay me for the privilege and it would come with more restrictions.

Re: Very interestingt

"The problem is that our developers make software for themselves and people like them. How do you get them to make apps for the common person?"

I don't think this correctly describes what exists. There are a lot of open source applications out there that do the same things as proprietary ones. Since you mentioned Google, I can think of very few things that Google has that there isn't an open source alternative to (mainly Search). People choose to use Google Meet instead of Jitsi or something like it because Google Meet doesn't require you to admin a server, do manual account management, or pay for bandwidth. People either don't understand the privacy differences between them or don't care, but the problem isn't that an open version is unavailable. The same thing applies to Docs and the online version of LibreOffice from Collabora, Maps and OSM-based software, and on and on.

Re: Very interesting

It doesn't run any interference. It isn't intended to. All it's intended to do is to get money from corporate users and hopefully return it to the developers, which I don't think it will actually do.

However, I am less concerned about the things that you mention. I already have a form letter for requests to add something to an open source project that aren't on my plan. It tells them that the suggested feature is not on my roadmap, that it is open source so they are free to contribute, and that I use a variety of methods to decide what is worth my time to implement it. Essentially, I tell them that if they want me to write something that I don't plan to write, either they provide me convincing evidence that a lot of people would benefit, they provide me with money, or they do the work. I will usually review their idea a bit so that, if it's something I won't merge, I can tell them ahead of time. If they have sent a request, they'll usually get something like that.

If they threaten me, they won't even get that. I do not respond well to threats. I haven't had the same experience that you have because I've received very few abusive communications. Those I have received have been from individuals, and I've informed them that I consider their messages inappropriate, and if they keep sending threats, add an email rule to automatically delete them.

Re: Very interesting

I think one central disagreement we have is this: "If you want to get paid to make Open Source". You may want to be paid to do all sorts of things which doesn't guarantee that you will get paid the amount you want. One of the downsides of writing open source is that others can take your work and use it without having to pay you. They might, and you can do things to encourage it, but you don't have any guarantee of that specifically because I have the freedoms to use it, modify it, and distribute it without having to do something for you or get your permission. The alternative is obvious: proprietary. You write the software, you sell the software, you get money. People who try to pretend that these two can be merged end up making something that looks and works a lot like proprietary software, then try to claim that they're operating in the same spirit as actual open source authors. They are not.

I don't think I'm telling you anything you haven't seen before, in fact anything you haven't argued a lot better than I have. For example, I can refer you to this article. This includes comments from the OSI and from you about claims of something being open source when it's not. I don't think you're claiming that this new license would be an open source one as it obviously isn't compliant with the OSD, but I think it's also harmful to try to claim it's even slightly close to the spirit. If you think the original open source idea has failed, fine, but make it obvious that this is a third, unrelated, path so real open source isn't damaged as badly if any of my concerns expressed above prove valid.

Re: Pay who first?

In the case of bankruptcy, payments in contracts are preferred over payments to stockholders. The specifics of the contracts would decide where the artists fall in relation to bank loans, landlords, server bills, or the like, but they would definitely rank above payments to the stockholders. Outside of bankruptcy, the payments to the artists occur first on the balance sheet, before the profit numbers are calculated, and any dividend payments or effects on stock prices occur based on those subsequent profit numbers.

Of course, this is not what we're talking about anyway. The discussion was not about when they're payed. It is about how much they're payed. It wouldn't matter if they were sorted above any other debt from the business if the payments were tiny, and the complaints from artists are not about where they fall in the sorting algorithm, but about the absolute size of the payments.

Re: Pay who first?

They do. If they were to go bankrupt, the artists would get paid first. The statement is misleading. I think what that was meant to say was that Spotify, being a for-profit organization, is supposed to figure out the bare minimum they can pay the creators of the music so that they have as much as possible left to return to the stockholders. Theoretically, a nonprofit entity wouldn't do that. Some nonprofit entities actually are run with their charitable point in mind and would fulfill that requirement. Unfortunately, we have lots of examples where that isn't the case. There are three models:

1. Nonprofit entity is founded for a specific purpose and run by people who believe in it. They collect a small amount for their labor in running the organization, a reasonable amount for organizational overhead, and the rest of it gets correctly spent on their goal. I've seen people who don't believe this exists, but I know several firsthand (volunteer, not executive). Sadly, they aren't the only kind.

2. Nonprofit entity is run by people who run it for the purpose it was intended for but mostly do it to collect money. They get nice salaries and plenty of perks paid for from the budget. Generally, the goal still runs correctly and gets funded sufficiently, but it could be better. For instance, ICANN.

3. Nonprofit entity is run by people who only want money and start trying to figure out how to extract as much of it now as they can. Whether that harms the programs they're supposed to be running, risks running the organization into bankruptcy, or violates the philosophy under which they were founded, they go ahead if it sounds like they'll get paid enough for doing it. The programs sort of run for now, but can break at any time when they either sell them off or run out of the funds to continue supporting them. For instance, Nominet.

Re: Very interesting

I disagree. I think it will be a harmful idea both in the short and long terms. There are several problems with it. I could write many pages on why, but I don't think you want to read that much, so here's a short list of the problems I see with it.

1. It limits the freedom to fork, because if I fork someone's project, who gets paid for use of my fork? Me or them? Or both of us somehow? This is an important issue because, if I get paid for the fork, the original authors will need to prevent me from forking to collect their revenue.

2. It introduces more incentives to get yourself included as a developer. How is the money allocated among developers? Does doing pointless work to increase your lines of code increase your payout? How much extra work does this add for the maintainers? If it is allocated equally to developers, then why don't I contribute a single commit from the accounts of everyone I know to get payouts?

3. It centralizes the authority for this in one organization. Nonprofit or not, that's no guarantee of anything. ISOC/PIR was a nonprofit, but that didn't stop a months-long fight to try to prevent them selling off the .org registry. Do we want to give much power to a single, license-mandated organization? Do we really think we'd have control over what decisions they make?

4. It goes against one of the aims of open source that caused Mr. Perens to resign from the OSI in 2020. From that article: "One of the goals for open source was you could use it without having to hire a lawyer. You could put [open source software] on your computer and run it and if you don't redistribute or modify it, you don't really have to read the license." This no longer applies for companies of a certain size. Sure, I'm not a company worth $5 million, so it's not me that's affected, but that doesn't make it good.

5. Your assumption that "an acceptance that 1% of revenue is a not unreasonable licensing cost". No, they won't do that. 1% of Google's revenue is $3.08 billion with more of that next year. That's just for projects that adopt this license. Other open source stuff doesn't get included. This means two things. First, if they pay it, they are almost certainly not paying any more to other open source software because they've already paid plenty in their mind. Second, they have a large incentive not to use this stuff and, where necessary, to build a different version. They have history in doing it. When Busybox demonstrated that they were eager to defend their licenses, Google made sure that they used something else in Android, and that got adopted elsewhere as well. Do you want to give companies an incentive measured in the billions per year to make sure these projects don't get adopted?

6. It eliminates the benefit of open source which made it usable without conducting license audits.

7. It continues the, in my opinion harmful, tendency to pick a villain and declare that their use of open source is exploitative. A license designed around my opinion on who is good or who is bad is not going to be a good one. Changing the person who makes the decision will not improve this. There is a reason why the original OSD had two explicit non-discrimination clauses and five in total. Encouraging people to drop this, which already goes against that definition, will likely result in even more of them being dropped.

I've long considered Bruce Perens one of the people who best expressed my views on open source, its structure, and its importance. That is no longer the case. If you want proprietary, write proprietary.

Not a Genius move: Resurrecting war hero Alan Turing as your 'chief AI officer'

Re: Erm...

As others have explained, there are some cases where there are legal rights to use an image or likeness which I'm guessing this company has not bought. However, the situation is more basic than that. Even if they do have a legal right to use such things, they should choose not to. Whether or not their actions can land them in a court and be ruled illegal, they are, in my opinion and I think those of others here, unethical, in poor taste, and bad ideas. You are allowed to do something that is all three of those, but it would probably be best if you didn't.

Re: What does the "C" stand for . . .

Of course it's more complicated than that. If I buy a car and it has a faulty brake system where, after a year of driving, the brake stops working and accelerates the car straight into whatever is in front of me, you don't blame me for having chosen that faulty car. The blame goes to the manufacturer who built it, and the consequences for them will be different if it's something they knew could happen or not. If I messed with the car and broke the brake, then it does become my fault. The software running the car is not something the average driver controls, meaning that there are plenty of reasons to hold the manufacturer, not the owner, responsible for failures that are entirely due to software malfunction.

The money that will be spent will actually intend to implement your decision, as it is the manufacturers who want to make sure that, if their software proves unsafe, they're not the ones who have to pay. There are situations where putting it all on the manufacturers is unfair as well. It really is a complex issue that needs more discussion and regulation.

Re: What does the "C" stand for . . .

Not really, though it could be brought up. If it relies on any laws from the original country, it is not a valid precedent, and if it is counteracted by any laws of the new country, it would hold no relevance. However, there are some cases where one common law country does consider common law decisions from other ones, so it might be considered if a similar case was brought up in the UK or US. There is a reason to think that courts would decide the same thing independently just because there's no logical alternative of what to do.

Elon Musk's latest brainfart is to turn Tesla cars into AWS on wheels

Re: WTF? This is meant to be a tech site - why is nobody saying...

El Reg hasn't gotten it wrong. There is compute capacity, and they haven't attempted to quantify it. We know that the total of all the computers in Tesla vehicles everywhere would make a relatively powerful parallel computer. Thus, there is capacity. The only problem is that the computers aren't Tesla's to sell, you can't actually use them as efficiently as you could a set of servers in a datacenter, actually trying to use this would put a lot of wear on parts of the car, the power usage would be much less efficient and would lead to problems from users who don't need to waste their electricity on this, and about twenty more things. So El Reg isn't wrong to say that there is a lot of compute capacity, nor are they wrong to explain why it isn't as useful as the raw compute numbers (which they don't have) would make out.

Google pulls RISC-V support from generic Android kernel

Given how many Chinese companies design chips around ARM cores, then manufacture them, then write the software that runs on them, I don't think that's the reasoning. China can, whenever it wants, stop paying ARM for licenses and still crank out as many ARM chips as they want. I'm sure American politicians will figure this out in time and make statements about how they'll prevent them from doing so which will have about as much success as their desires to prevent them from using RISC-V, which is about as much if they announced that they will be banning them from using iron.

Meanwhile, to keep the code in the project means plenty of developer time spent testing and fixing these things, all for hardware that doesn't exist. It's not surprising that Google wants to wait until their work is achieving something before they take on the effort of maintaining it. Almost certainly, if Qualcomm starts to make RISC-V phone chips and gets someone to build a device around them, they will merge plenty of changes to account for how their version differs from the theoretical version. When they're done with that, Google can merge the important ones back into their project rather than trying to do that continuously. The effect of the delay to end users will be unnoticeable.

Software support chap survived breaking his customer

Re: Don't anybody move!

True, although the other clue that this probably wasn't DOS was that the temporary data was 9 GB of data covering 11 years of business. While it is possible to have a DOS machine with 9 GB of storage space, it's not very likely and probably wouldn't have been a single directory containing the lot.

Re: Picking the data to delete

"It was "courageous" (as sir Humphrey would say) to use an evanescent MFS file system as enough not entirely clueless users would place files in subdirectories of /tmp and run their own periodic job to "touch" those files so they escaped the grim reaper."

Or it was the necessary condition to educate them that temporary directories are actually temporary, not just extra storage provided if you thought to use it. It had to be pretty obvious that the quotas on their home directories existed for a reason, namely that disk wasn't unlimited. Using a different directory for temporary data was acceptable, while trying to use it to bypass the quota was not.

I often create a ramdisk on my systems because I have processes that generate gigabytes of actually temporary files, so storing them in memory makes the operations on them faster and doesn't wear my disk storing data. I know that if the computer loses power that the operations on those temporary files need to start again. That's why the ones I put there are truly temporary. Losing data there is a good reminder that it isn't intended for important files. That's what I have a nonvolatile disk for.

Re: Don't anybody move!

"Percy was using remote access."

Hence people considering that it probably wasn't DOS the computer was running. Remote access was the basis, the OS in use was what they were trying to guess.

Re: Picking the data to delete

Don't count on that. On some systems, /tmp is stored in RAM. Disconnect power and /tmp is wiped. I write software to assume this will happen, which means if I need something to recover from a crash, it has to go somewhere else, such as /var/tmp, although a dedicated location for the program is more likely.

Encrypted email service files DMA complaint claiming it vanished from Google Search

My assumption still involves them being wrong, but assuming that the problems they encountered were due to their domain being blocked rather than their server being blocked. Either that or they actually do encounter sites similar to the one I described a lot more frequently than we do. After all, I can't say that most sites accept my domain, just that most sites I've used have accepted it. That's a rather small subset. Either way, I'm not sure that they were intending to mislead rather than simply being incorrect.

One possibility is that their experience differs from yours. I've described my own history with self-running a mail server in another comment, which involved trying to get off some spam lists that I was on before turning the server on for the first time merely because of my IP address. The part I didn't mention was that, before I did that, I had to get a new hosting provider for the mail server because the company that I used to host other things was on a lot more spam lists. The ability to quickly set up a new rented server was great for me when I wanted to host something, but it was also useful for spammers. I didn't even bother setting one up there because I checked on it, but if I had, I think it would have been rejected by a lot of sites using expansive blocklists.

Maybe the person commented as they did because they had such a mail server, either one of their own or run by a company that were bad at their job. They may be overgeneralizing how common it is, but I'm inclined to think that they actually did encounter the problems they're describing rather than making it up. Even now where I have email through my own domain, though unfortunately not on self-hosted instances, I have encountered at least one service that rejected the address. I was registering for a promotion, and perhaps they were afraid that my domain was going to try to register a lot of times, although they had no problem with Gmail accounts and setting up a bunch of those is frequently managed by spammers. Either way, they did not accept it. This has only happened the once, but some sites are that aggressive.

Re: Talking of blocking emails…

That is the challenge of running your own mail server. You have to be careful not to have even the smallest misconfiguration from unofficial standards, or you will end up on spam lists that are difficult to get out of. Still, you probably have a chance of doing that. The other problem is that many of those lists will still list you if someone from an IP range even slightly similar to yours does anything dodgy, including if they are hacked. It doesn't matter that their IP address is different from yours, attached to a different account, and your addresses have been operating under your control and without issue for years. Someone will still block first and ask questions later.

I have operated my own mail server before, and with relatively good results*, but nowadays, I let someone else do that hard work and just point my MX records to them. I don't like having to do this, given that I self-run all the other services attached to my domains, but if I need the emails to definitely arrive all the time, I don't have much of a choice. It might be different if I had a dedicated IP range for my services which I could keep clean, but I don't. I do wonder how aggressive this will be when IPV6-only servers eventually happen, because I am able to get my own IPV6 block more easily than my own IPV4 one.

* To my knowledge, my self-run server didn't get added to any spam lists. I was certainly able to send from it to Office365 and GMail. However, there were some smaller lists that included massive IP blocks which included my host. I had limited success applying to have my IP removed from the blocks. I knew it was only a matter of time before those seemingly unused blocklists would end up affecting something I cared about.

Shouldn't Teams, Zoom, Slack all interoperate securely for the Feds? Wyden is asking

Re: Contradictory statements?

Why are those contradictory? The solution seems simple: encrypted backups. If you have the keys and access to the disks, you can play them. If you intercepted the communication, you can't. Ideally, the backups would be stored by the user so the encrypted versions can't be obtained by the provider. It would also work if the provider stored the backups but not the keys, because they would have access to the files but not the ability to read them.

That's used by all sorts of protocols. My messages are end-to-end encrypted in transit, but nothing stops me from keeping them on my device once I have them and copying them onto another system for backup.

Microsoft dusts off ancient MS-DOS 4.0 code for release on GitHub

Re: Ancient!

They're not suggesting that you install it on your work machine, you know. It's there to look at, run for nostalgia's sake, or if you somehow have a use for it. Otherwise, you are free to ignore it.

I'm sure someone would do something, but less sure that whatever they did would be useful. Refurbishing hardware that's fifteen years older has a lower cost to result because they generally want it to boot up the same environment that it came with. They may have to emulate some hardware to do that, but the interfaces aren't complex, and they're not trying to bolt too much on. For a modernized Windows 98 to work, they'd have to do a lot more, such as getting it to boot natively on modern hardware, to interact with modern peripherals, and to run the software needed to perform modern tasks. None of those things are straightforward in comparison to getting an old computer with most of its old parts to run the software it came with.

The hyper-clouds are open source's friends

Re: very selective choice of open source

Of course there will be many projects that they used, not all of which they became major contributors to. However, that's not enough to make them the bad guys here. Do you have Linux machines? I'm guessing so. How many kernel contributions have you made? If the answer is zero or few, does that make you a bad person? I don't think any user, even if they're large, have a moral requirement to write code or donate money, and if you really want to make a list of users who don’t do those things, the cloud providers are not going to top it if you sort by number of projects used without support. Open source developers know that people will use their software, including making money, without being required to donate back. If a developer isn’t happy with that, they are free to write proprietary software instead. The licenses that implement this are not ambiguous about this.

TikTok ban could escalate US-China trade war, ex-White House CIO tells The Reg

Re: As he peered down from the clouds

"Unfortunately the most prosperous and advanced periods of humanity follow.wars where for a generation or two we ask why we did it,"

I don't think that one's supported by history. Maybe you're thinking of the prosperity that happened after World War II, except that there was another war including 16 of the players just five years later. It was smaller, but it still involved plenty of dying among those participants who had just finished an earlier war. If you're thinking about post World War I, you do have slightly longer without another one, but the prosperity didn't last very long and it wasn't the start of another war that ended it. If we consider farther back in history, one long multi-party war in Europe usually degraded into smaller wars with relatively small gaps in between.

It would be nice if we did have the aversion to war, even temporary, that you describe, but I don't think we do, and even if we do, it wouldn't stop us from still having to get into wars when someone else starts them. As we have also recognized, ignoring something because you don't want to have to fight usually doesn't prevent you from having to fight eventually.

Re: Who will actually lose out

People have lots of theories. They say it's about social media not wanting competition. They say it's about some content moderation decision. They say it's an attempt to affect the election. They say there is a real propaganda or privacy risk. I don't think any of these really do it.

Politicians have decided that China is dangerous, and this is a thing that a lot of people use which was created in China, so it must be dangerous. Exactly how it is dangerous and what you could do about it that would best reduce the danger are problems politicians don't want to bother with. Whether this law is legal is another one they didn't spend much time on. In their minds, it is a threat and they've done their duty in responding to it, and that's the level of detail they care about. It could actually be a serious threat, it could be completely harmless, or it could have potential problems that are being exaggerated, and it wouldn't matter to the people passing the law. They probably are thinking about whether they can use this as an argument when campaigning, but even this is not their primary concern.

UK's Investigatory Powers Bill to become law despite tech world opposition

Re: Harvest Away......Avoidance Is Possible.........

If they're looking, they would just run a query on the history: SELECT phone_number FROM calls WHERE name_known=false ORDER BY call_frequency;

There is your list of people who use burner phones. At the beginning, that's a lot of unknown callers and the query isn't useful. If they start attaching names to numbers, the list gets smaller and only consists of people who are intentionally preventing them from making those links. That is your list of people of interest.

And yes, there are ways to fight against that, just as there are for every other abuse. Those methods are more difficult, easier to mess up if you think you know more than you do, and more easily made illegal when only a small, easily scapegoated subset are doing them. For example, if the search is for numbers without names linked to them, you could get a phone under someone else's name, but that requires more effort and could become challenging to keep up, not to mention that another query could identify people with multiple numbers that behave in completely different ways (calling different people from different locations). Turning something like this into a 1984-esque infallible surveillance regime is nearly impossible, but turning it into something pretty bad is much easier.

Re: Harvest Away......Avoidance Is Possible.........

If enough people do that, they can always come back and ask for another law. You can use a burner phone now, but there are lots of countries that forbid them and require you present identification to have a phone number. They're not universally dictatorships either, though many dictatorships do it. Of course, that doesn't mean that every phone in the country has done that, but it does make it harder. They cannot make it impossible to get around some of this, but they can make it difficult unless you have plenty of knowledge and a lot of determination. The other problem is that if you think you have plenty of knowledge and don't, you could end up standing out like a spotlight. Just because circumvention is possible doesn't stop this from being a bad law.

Ring dinged for $5.6M after, among other claims, rogue insider spied on 'pretty girls'

Yes, I mean the customers. I was replying to this sentiment:

"It astounds me how anyone (never mind that it is so many) thinks that cloud-connected CCTV on/in their home is a good idea..."

Sometimes, I feel the same thing. I don't want to use a thing and I just think that everyone who does must be wrong in some way. Telling people that they're wrong won't get anything done whether they are or not, and in most cases, the people who want a thing I don't have a reason why they wouldn't prefer the alternative that I do. If I want to change their minds or even understand why we have this difference in opinion, I have to look at it from their perspective.

The company's perspective is much easier to understand: they want money. Adding security to the code costs developer time, which means money. Adding security precautions to collected data slows down the work of anyone who has to access it, which costs money. Reviewing those who abused it would take time and might result in firings, and then they'd have to hire someone else, and that costs money. So they do as little of those things as they can.

BOFH: Smells like Teams spirit

Not really. Enter will nearly always send a message, and you use shift+enter to add a new line. There is an exception when it decided to make a list for you, but then you have to press enter to get out and again to send it. So basically, users probably aren't doing the multiple message thing by accident. I think they do it because, if you have to write a large paragraph, sending it in pieces means the recipient doesn't have to wait for you to type the whole thing to read it, but that's only of use if they're already in the conversation. Otherwise, it makes them want to avoid helping you instead.