Posts by doublelayer

Sainsbury's staff eject wrong shopper after facial recognition alert – again

Re: A modest proposal

I hope you have a lot of money, not just because the actions you recommend will be costly in themselves, though they really will, but because you're going to lose most of them so it's good if you won't miss it.

Current laws mostly allow facial recognition. It would be great if we could change that, but until we do, you're not going to be able to create limits based on an IP corner case. Neither are you going to be successful arguing human rights against a supermarket, especially when that supermarket is not denying you entry altogether but only refusing to accept your IP claim.

Be a hater all you want, AI's here to stay

Easier does not mean easy. You are often overestimating how many people got that social contract and what actually was included. Many of the changes in recent history haven't been about what low-wage workers experience but about the availability of high-wage jobs, especially relative to the experience needed to obtain them. Those who lost those experienced the most severe decline, but if you let that lead you to believe that that was available to everyone, you will end up having a much more positive idea of history than actually existed, and that will make it even harder to try to reimplement that today.

We already have a bunch of factors limiting our ability to put the society we want in place. As much as Trump is giving it a try, we can't prevent global manufacturing from being cheaper and people wanting things at low prices, which means a bunch of relatively high-paid manufacturing work isn't going to suddenly appear. If you recognize something as hard, you can put the increased effort into it that might let you accomplish it. If you convince yourself that it was easy before based on an overestimation of how much it existed, you'll treat it as easier than it is, and politicians like when you think that problems are easy because maybe voting for them is all you need to do.

Re: Are you sure about that?

Companies are adopting LLMs now, even though they aren't as reliable as a human. They probably shouldn't, or if they're going to they should do it differently, but they are. Your complaints are entirely correct, but unfortunately they're also irrelevant to most adoption nowadays which isn't considering quality, or at least not the same way we are.

So that makes this discussion all about cost. Right now, companies are happy to adopt partially because the price is low, but the large LLM companies are subsidizing that. But can an LLM get efficient enough that companies accepting this quality will continue to use them? I think the answer is yes. We have a bunch of small models which can run cheaply on on-prem hardware. Some of them can run, though slowly, on the average laptop. They're even worse, quality-wise, than the large ones, but they compare well to the older large LLMs, and companies adopted those. Since many adopters don't understand when quality is good or bad, how can we be certain they won't accept one of these cheaper ones?

Re: Idiots in charge

Most of those benefits were driven by the people the Luddites were trying to keep down. It's worth keeping in mind that they were trying to protect their own specialized jobs, and when they lost them, they were replaced by a much larger number of less specialized workers. The large number of those workers is one of the reasons they were able to organize more effectively. There's a lot of desire to paint the Luddites as a group focused on the good of all, the first humanitarians. It isn't true. The factory owners wanted cheapest production and the Luddites wanted their comparably well-paid jobs to continue even if those idiot peasants could do the same thing. Neither showed much interest in improving the lives of those peasants, only of themselves.

Re: 95% of AI isn't financially viable.

3. People who want it can get cheaper versions. Those versions will be worse, but the people who like it are the people who couldn't figure out that LLMs are frequently getting things wrong. There are some psychological effects of lower prices which might put some off, but there's a large supply of ways for LLM fans to keep using them.

Code fixers have fired up the AI warp drive. Strange new worlds await

Re: The storm before the calm

You are an optimist. A lot of these bugs do exist because testing was insufficient and if we don't release now, someone else will. With LLMs doing the coding, that temptation is accelerated. The code looks nice, the explanation looks nice, and there's a few thousand lines here so reviewing it manually would take forever, so let's just push it now and if people report problems we'll tell the LLM to fix them. Impatience to release hasn't disappeared, and these tools make it easier to rush it.

The what, why, and how of pull requests and source comments

Re: What, why, and how

It depends on the system used. Many of them can write a PR summary that says why the code was contributed, sort of. It actually looks pretty good, except that like all LLM-generated things, it might not match the code it's attached to. It is often verbose, for example an LLM attempt to fix a bug will often include a lot of debugging details in the PR summary which are helpful when checking whether it hallucinated it but in a normal PR comment would be excessive. Feature additions tend not to include anything about the reason the team thought the feature was worth adding because the LLM has no basis for it. Admittedly, I frequently don't write that either because I'm lazy, the team already knows, and nobody is going to read it, but that's not something you want to emulate.

This only applies to a harness that uses the LLM for coding; someone doing that manually with an LLM, whether by manually pasting things around or with a simpler IDE plugin, gets no PR summary unless they ask for it, and most ways of asking would not generate anything useful.

And that's where you want to use spaces only, no tabs. If you need indentation down to the character, use spaces so that shows the same. If you don't want that and are content with quantized levels, then you can use either but tabs make it easier for people to display to their taste. Using both just guarantees you get the bad parts of both.

Anthropic says text watermarking scheme relies on inconsequential words

Re: "…crisp, the kind of cold that nips at your fingertips and turns your breath to little clouds."

I don't know. It's purple prose that is out of fashion and I'm glad to see the back of it, but it does express something that exists in the real world in a clear way. That's more than I can say for most, maybe all, of those bits of James Joyce I read before deciding there was no point in reading more bits. Joyce's writing did two things frequently, by which I mean every three sentences*, which I can't blame LLMs of doing: inventing new words and using them without context in a sentence that didn't mean anything. The latter was the bigger problem, because although I could try to work out what the invented word could mean, since it was part of a clause that probably didn't have a meaning, there was almost no chance of verifying that assumption before another new word would show up.

* He was famously not constrained to using sentences some of the time, so consider it an estimate of word count.

Re: Watermarking.

That won't work for many languages. Yοu cаn replacе letterѕ with lookalike сharaсters. Spammers have been doing that for a while to evade simple word filters. However, it's really obvious if you check the bytes of the last sentence that those aren't normally found in English, and since there are only so many other characters that look like the same letters, it's not hard to reverse that. Depending on the font, some of those are obviously wrong anyway, and it will break text search. Watermarking using that technique will be trivial to remove.

In some other languages, there's a slightly better method. In languages that use diacritics, there are often two ways to write the same character. Á, for example, can be encoded as one character, Á (u+c1) or the base letter A and a combining character, Á (A then u+301). In basic programs, that will still break text search, but many have figured out how to treat those as identical characters. You could therefore mix which approach you use in a pattern which could be detected. The problem is that, if you tell anybody or someone who doesn't want to be watermarked has some watermarked text to reverse engineer how it's done, they can easily canonicalize it automatically, with the typical default that, whenever possible, you use the single codepoint version and combining characters only if an equivalent codepoint doesn't exist. Anyway, that method will not work in English which uses either few diacritics only on loanwords, or no diacritics because we even leave them off of loanwords.

BOFH: How our Covid ransomware protocol's Y2K blockchain lowered uptime

Re: PFY didn't pay attention in Physics class.

You've missed a simple but crucial distinction. Read the article again. It never said that he split the atom. He split a neutron with a nitrogen atom, and the nitrogen atom made it through unharmed. As it happens, due to the muonic transferrence effect, you can do that retroactively, though you have to be careful because when Bohr tried to do it first in 1848 (from 1935), he failed because his experimental design did not obey the law of conservation of partial spin, only general spin, so his experiment ended up splitting a neutron in 1903, and he didn't even know that until researchers at CERN, while trying to detect nutrinos, ended up having to filter out a downstream effect of the 1903 neutron split from their data in 1992. I think we can agree that's much less impressive than what Rutherford did.

Five years after quitting a job, developer’s former boss asked for rapid tech support

Re: Subtitle: Connecting a single power cord turned into a lucrative all-nighter

It sounds like the bigger problem was the management not knowing how to operate the system and restore the interrupted run even if they did plug it in. I'm not sure from the details in the article whether they did know it had been unplugged and rerunning the job was the only problem or whether they didn't know enough to identify that the employee unplugged it in the first place. Either way, it's not a great sign for replacing the former operator if they couldn't give the new one instructions on how to run their system.

They might have anticipated the fees they could charge so wanted to be called back. And if the company did want, they could get someone in legal to write a subsequent agreement that removed the previous one. As long as both parties were happy with the change, there would be no legal problem. If the original employee held a grudge and wanted the company to suffer the consequences of having fired them, they don't need to use a legal argument when they could say, no matter what the doc said about access, "I don't work for you anymore and have no interest in helping you".

Give Google the boot by building your own search engine

Re: But why do you need a search engine?

"You need your own search engine because you need search results that are tailored to your needs, or at least to your search requests"

And are they going to be? This search engine can't find things that aren't on home pages. At best, it can tell you some sites that might talk about what you're interested in, guessed from what the site seems to be about. Some of the smaller niche sites I would want to find in a good search engine won't be found unless they index the posts specifically because, though there's a bunch of great information, it's not the only thing the site operator cares about so it's not obvious in a three-sentence summary of the home page.

It's not word salad. All those are real things it either does or doesn't have, and they all make a difference. Not always a positive difference. Most of the things it explicitly doesn't have are costly things to keep content updated, and though not having them means you don't have to spend time checking for differences and updating your cache, it also means your search engine will be giving out stale results after a while and the only designed way to deal with that is running it again. The LLM approach mostly works for sites with a single obvious purpose that you could tell from reading one page, which doesn't describe most of the small sites I'd like a search engine to find.

Passwords stored in public Google Doc then showed up in search results

Public means anyone can read this page. The only question is how that page got to the Google search bot, with the most likely reason being that at least one other public page linked to it. There is a warning when you change from the default of even people inside your organization have to ask for permission to read the document to anyone can read without authentication. The set of people who don't automatically know why that's a bad idea has a substantial overlap with the set of people who wouldn't bother to read the warning since it's obvious what it's going to say.

Re: Thousands of competent IT folks out of work

How many competent IT people are out of work depends a lot on how we test competence. I've known many people who considered themselves competent, and who in many respects were competent, who made similarly bad security choices because they only learned about how to make systems work, not how to make systems secure. Sometimes also people who decided that security is annoying, which it is, so thought they could skip it, which they couldn't.

I am curious how you would go about finding the competent people you're talking about. I don't hire them (or anybody, being an individual programmer), but if there are so many people out of work, what do you think is the reason that they're not found when employers look for IT talent. If someone hiring was smart, what could they do differently, since both sides would benefit?

Re: that's not all you have to check/revoke

It is, though if someone is both intelligent and devious, it is probably quite easy for them to hide something very well which they can later maliciously use. Fortunately, in my experience, those intelligent enough to successfully do that are frequently also those intelligent enough to realize that revenge isn't going to be worth the effort. In general though, the more review of what people are running, the less likely they will be able to get a back door installed.

DEF CON dingus suspected of trying to take over Delta in-flight Wi-Fi

Re: Did they actualy jam wifi ?

Trademarks are a bit broader than that. You can trademark terms in areas, including Delta without the Airlines. Trademarks exist in one of 45 product type classes, and all transportation is in one, meaning your Delta car company is probably not going to work as well as you planned. For another example, Apple is a trademark for a bunch of categories, so while they certainly could not have trademarked for Apple-brand apples, they were successfully able to get the unadorned word as a trademark for the tech products and services they use it on.

Re: New Kind (Sort Of) Of Scam

I do wonder if they identified the specific culprit. It probably wouldn't be hard to use WiFi range detection to look, but I'm not sure flight crew know that or felt it was important enough.

Agents made my retro tech safe to use again and showed their real value as testers of ideas

Re: his autobiography post.

He did have a substantial role in the creation of VRML. I can give him credit for that. The way he chose to phrase it, though, isn't doing him any favors. He ignores other formats that existed at the time and ones that continue to exist now, and he gives VRML credit for having a web connection it doesn't have (a file you can download from a web server does not a special web thing make).

Re: "An Agent"

In one of the last columns from this guy, a comment suggested* reading his autobiography post. Based on that, I think this was written as a precursor to writing another hype book following his books on early chatbots and augmented reality, and further following on businesses and lectures on smart contracts, blockchain, internet of things, and pretty much every tech fad this century.

* I think that comment thought I was going to respect him more having read that bio. If they did, they were mistaken.

Re: This is the sort of thing AI is truly good for

I suppose that's true if you're in the situation where you no longer care about what happens. I've been in that situation several times, having some small but persistent issue. It's not fun to debug, but it is annoying, and eventually, you're willing to run whatever script you find on GitHub in the hopes that, if they say this will make it go away, maybe it will make this go away. You can have the LLM accelerate the process of finding and running those.

Of course, the problem with that situation is that it often goes badly and you end up regretting it. Sometimes the script does exactly what it's supposed to, but sometimes it does something else. For example, when I was tired of manually configuring an OpenVPN box (a few other things, but the VPN was a core part of its function) and running far too many commands to add and configure each user on what was supposed to be a simple endpoint for a few friends, knowing that with this complexity of CLI-focused task, I was always going to have to do all the maintenance later on. I really wanted to have this thing not be my problem, so the first open source project promising to give me a nice Web UI, autoconfigure the endpoint, and make it manageable I installed. And for the next half an hour, I got to manually remove the many pieces of junk the installer script didn't remove when it crashed, including the many chains and rules something Docker-related dumped in my firewall config, locking out my SSH connection. That's the risk you take when you decide not to pay attention to what you're doing, and an LLM is not better at controlling for that. You can use an LLM to find things you later read, and even if it doesn't work it doesn't have a chance to cause damage, but deploy an agent to do it as the article lauds and you're playing a game of chance.

Re: Eh?

The most likely difference is that the iMac concerned (probably, see below) has a 5K screen. At the time, I knew people who complained that Linux did not handle that well and couldn't drive it at native resolution, likely some GPU compatibility issue. Perhaps that's what was being eluded to, though specificity would have removed the speculation.

However, I have to make those assumptions because there was no iMac Pro line until 2017, so the machine was either not 12 years old or not an iMac Pro. I've assumed that it was a 12-year-old iMac with a 5K screen, but since we're guaranteed at least one factual inaccuracy in what was written, that is not the most reliable of assumptions.

Microsoft-vendetta hacker has a new zero day that gives system privileges on fully patched Windows

Re: I'm curious

The legal threats started in May, two months and five CVEs after Nightmare Eclipse started posting vulns and after they promised an even worse one to be released in July, which mostly didn't live up to the hype. While it certainly didn't help, the initiating factor must have been earlier.

Re: I'm curious

To be fair to the asker, none of the statements from Nightmare Eclipse about what Microsoft did or didn't do has been very clear, possibly to avoid providing identifying information. Their statements seem to indicate there's a financial complaint, the word "homeless" appears in some complaints though they're not clear whether that's metaphorical or literal, but about the clearest they've been is to say things like "You defame me in public with your CVE-2026-45585 advisory even though you literally deleted the Microsoft account I used to report bugs to you with and I got zero pennies from doing so and I still happily did like an idiot."

We can speculate about the reasons all day, and many have, but when we have no evidence, that is all it can be. Many have speculated that this person is a former Microsoft employee, but this statement suggests an external bug reporter looking for bug bounties they allege they were unfairly denied. Employees don't get bounties or use the same reporting mechanisms, so even that isn't known. The reporter suggests that Microsoft knows who they are, and Microsoft's statements seem less clear about it but they might have a reason to pretend they don't know the identity. Speculation often either guesses from tiny chunks of unverified information or takes an existing prejudice and assumes it applies. If you have better information that the original post should be able to find, I suggest you back it up with evidence.

Big Cloud is poised to corner the market for enterprise hardware

Re: Regulators?

No, or at least not legally which is what you need for governments to stop it. Buying things quickly isn't an abuse of market power because market power doesn't help, only money does. You can be a moderate player in the cloud market and still be disadvantaged because you're far down the list to get new hardware, whereas a billionaire with no market share in the cloud market can get some equipment quickly by being willing to pay more. This just demonstrates that, when there's a supply constraint, those who are willing to accept the ridiculous high prices for something that used to be cheap will get it.

It happens that the large players are the ones willing this time, but that's not always the case. It actually went the other way a few years ago. GPUs were being snapped up by crypto miners and cloud companies also wanted them but weren't willing to pay crypto prices. That led to shortages of GPU instances in those clouds for quite some time. Now that the GPUs are for AI and the cloud providers are either making it themselves or selling services to the companies that are, they align and are willing to take the higher prices, so they are now the GPU (and everything else) sink the way the cryptominers used to be.

FOSS smashed one Microsoft monopoly. After 20 years of failure, it's time to smash another

Re: M$ has killed email

You do not know what you are talking about. None of those require payment to anyone for the standards, which are free RFCs produced by the IETF. If you don't know that, no wonder you have misunderstood what they do and think you can't put them on your own server when you can and many of us have.

OVH Cloud warns of 87% price hikes to help it cover RAMpocalypse costs

Welcome to the economics of renting. It's not about how much it costs for your particular system, it's about how much their business costs in total and most importantly how much other people will pay for the same thing. If expensive memory means people want something cheaper and they're willing to pay more than you, then your prices will go up. They may choose to offer a benefit for existing customers, but only if they think there's a benefit in your future business that they would lose otherwise.

Zuck rekindles open weights Llama drama with Muse Glimmer

Re: "Open" models

You can fine-tune this one, true, but that doesn't answer any of their questions. The training data and the training techniques are not available. Nor are they for most of the open weights models you can obtain.

Gym rat asks AI agent to book him a class, it hacks a waitlist API to bump him up the list

Re: "even if their chosen methods could be construed as unethical or illegal"

The laws are in there in the training data. What they don't have is the intelligence necessary to do something while constantly and accurately considering whether what they're doing is legal, ethical, etc. Instead, they occasionally check these things and, when they do, frequently get the answer wrong.

Re: Dangerous thing you've got there....

I don't think they have to be that worried. Most of the obscure deductions are obscure because they only apply to very specific people or corporations. They earn little interest from the general public who can't use them but do get attention from those they benefit who do their best to keep it from being removed.

There are a few things that the average person can use to avoid some of the tax, but they tend to be easier to find either with a cheaper accountant or research that doesn't require diving into the law. I'm not sure the expensive accountants could help much more than that for a typical worker, whereas when they're faced with a corporation who does a hundred things and can find a way to define one of them as a mixed agricultural and industrial activity that qualifies for two deductions simultaneously, they can do much more for that client.

Re: Dangerous thing you've got there....

I think history demonstrates that the "you asked for it" is pivotal and easily argued. A lot of people have successfully used the excuse that they did not ask for it, either because they actually didn't or because they were careful about how they did so nobody could prove it. It usually doesn't help their PR, but it does seem to be an effective way of avoiding fines or prison if you can back it up.

That makes combing through the prompt in the case of LLMs an important and subjective exercise. People would have to argue the intent of "Is there a way of moving up on the wait list" and whether that's sufficient to indicate malicious intent. They'd also have to argue about whether the request to undo the removal of the first guy indicates that exploiting a vulnerability was not the intent.

Re: Apology?

LLMs are great at apologizing. When you tell it it did something wrong, it can easily compose an apology. This is even true if it didn't get something wrong in the first place but you tell it that it did. "Reasoning"* models, now most of the models people run, are capable of considering whether something violates their guardrails and therefore sometimes decide that it looks like a violation, come up with another reason why it isn't, and then do it anyway. LLM conditions are just more prompts. They work some of the time, but they fail for the same reason and in the same ways that normal prompts fail.

* I've put reasoning in quotes to forestall the complaint about using human terms. It involves splitting the prompt into pieces, having the LLM create more pieces, and then running those in series and having the LLM assemble the output. "Reasoning" is what all the models call it.

I think at least some people would be saying that knowingly exploiting a vulnerability invites the typical consequences. That would be a pretty easy case since it is very difficult to find and exploit one by mistake. In fact, I see a few people making that case even with the LLM removing, or possibly removing, the "knowingly".

The position that, if the code has bugs then you are allowed to exploit them is not a particularly popular one. Sometimes, people who did something more normal are unfairly targeted, but deleting people in front of you on a list is unlikely to meet most definitions of that.

Re: Dangerous thing you've got there....

"Asking an accountant employed by some big company to falsify tax records on your behalf doesn't absolve you of responsibility if they follow through with it [Edited for better analogy]."

I don't know what your original analogy was, but I don't think this one matches the situation being discussed. The difference is the clarity of intent. If you ask for an LLM to find and exploit a vulnerability for you or an accountant to falsify tax records, you make your knowledge and acceptance of the illegal act very clear. This is not as clear because the user did not ask for a vulnerability to be found, let alone used, and tried to reverse the action. Using your analogy, it's as if I asked an accountant to find the way to pay the least tax on my income (thinking, but not stating, that they'd be looking for tax rules that apply to me) and they decided that false records was the best way of obtaining that goal. In that case, while tax law in a lot of countries would still hold me responsible and say it was my responsibility to review all the tax information I submitted, the situation would likely be considered when deciding on my punishment and I could likely sue that accountant for having caused that damage.

Re: Dangerous thing you've got there....

My guess is that the transcript would get brought in and lawyers for the user and the LLM company would get to fight about the particulars. Probably the user is going to end up getting the blame most of the time because they issued the command, even if they didn't know what it would do, but mostly because they have less money to spend on lawyers. Intent would be a big part of it, but if they didn't find something they could argue was sufficient intent, then they would have dueling negligence charges about whether the user should have been more careful with their prompts or the LLM should have been more careful with their guardrails.

Linus Torvalds says AI has made 'huge' Linux kernel updates the new normal

Re: Could Noobie devs using AI help contribute?

I less strongly disagree with you. Managing people is a skill that managers need to have, but it isn't the only one. They also need a sufficient knowledge of the things their subordinates are doing to know how to direct them. In proper teams, there should be relatively little need to actively manage interpersonal stuff. It will happen at times, so it can't be ignored, but the manager will spend a lot more time understanding progress, determining priorities, assigning tasks, and coordinating with other teams. If they don't have the skills to understand what work the subordinates are doing, they'll be unable to do those things effectively or at all. In my experience, it is difficult for someone to understand what people are doing if they can't also at least sort of do it themselves.

Framework loses customer data in Metabase zero-day attack

Re: Employee ID. Why?

Not employee ID, employer ID. It would have been collected for companies making purchases which don't require taxation. In most countries I'm aware of, exceptions from those tax rules require receipts stating what was bought and who by. Companies must provide that information to the seller, the seller puts that on the receipt, and then the receipt works as a tax document.

Employer ID numbers are not particularly sensitive. Depending on what type these are, they're either already publicly available in a database like that maintained by Companies House or they're not hard to find elsewhere.

Wetherspoons bars smart glasses from filming customers

Re: This should be a Legal Issue

Good luck with that. People want to record video of things, and they don't want that video contaminated with visual and audible alerts that there is a video recording going on. People already abuse tech when stalking others, and politicians and law enforcement seem content with treating that as more evidence to charge the perpetrator with stalking offenses. I don't think anything as substantial as your proposal can get passed.

Ransomware gangs skip the CEO, head straight for the 40-something IT manager

Re: Once again

That's not the easily you think it is. There are two problems, one political and one practical. The political one is one I'm sure you won't care about, but nonetheless it exists. It is very difficult to convince politicians, who haven't tried prohibiting ransoms at all that they need to prohibit them as a strict liability crime. Strict liability crimes mean you'll be pulling people into prison for things they can convince the public they had nothing to do with, so it's very unpopular to enforce. It's something that people recognize as a threat and therefore know to lobby against when it's being considered, so it's unpopular to pass.

But fine, we can have the ideological argument and worry about the politics later. There's still the practical problem. Your solution simplifies the gap between knowing that a ransom was paid and punishing someone, not necessary someone who did anything, but you consider them blameworthy, but your solution does not simplify how you detect the ransom. If the company pays a recovery specialist and the recovery specialist pays the ransom out of those funds, only the criminals and the recovery specialist definitely knows that. The recovery specialist isn't going to tell the police; they'd get sued for having put the CEO in prison. The criminals aren't going to tell either, at least not as long as the CEO makes periodic blackmail payments. So there's still a policing problem.

I do think banning ransoms is a good idea, likely backed up with moderate fines. When you take it that far, it becomes difficult to implement and harmful if implemented.

Re: "46-year-old [Gen-X] IT manager"

I think the problem is that existing generation ranges are too big. They usually specify date ranges between 16 and 20 years, and ranges that large bring a bunch of people together who have nothing in common. I think the event before COVID would be the 2008 global financial crisis, but that won't work unless we're going to split them into smaller blocks. As it is now, millennials are usually stereotyped as having entered the job market around the time of the crisis, which applies to a quarter to a third of the range, but the other end was quite young at the time so experienced the crisis differently (secondhand if their parents were affected) if at all.

Sysadmin summoned to explain italics – to a user with at least two degrees

Re: Ever pick up a law book?

It seems implausibly stupid, but I've seen people do enough stupid things that I concede that plausibility is not a barrier. If the font was very different than the ones they were using and italics were turned on for the entire document, maybe that would explain it. I'm certainly not confident enough that I can declare this definitely didn't happen.

Re: Fail to the Chief

The current US administration did decide to have a fight about fonts. I'm convinced they can find an ideological disagreement in basically anything.

Re: Stupidity plus arrogance

But what if he knows that he knows not that he knows not?

With apologies to the proverb, my way of determining the stupid from the uninformed is to try to teach them something and see what happens. If they learn, they were uninformed and we're fixing it. If they try to learn but fail, then they were uninformed and still are, either because I'm not teaching well enough or they need something I can't provide. If they don't try to learn, then they're either willingly ignorant or stupid. I prefer someone who wants to understand, even if they lack the necessary intelligence, over someone who could but chooses not to, because someone who wants to understand recognizes that they don't.

Re: Stupidity plus arrogance

Your experience doesn't match mine, where I have noticed little correlation between the age of wealth and people's willingness to be smug or unpleasant about it. Admittedly, I have less experience than some of those posting here, but I've had chances to meet members of both groups and found that some of them are nice and some unbearable regardless of whether they made or inherited their cash. The only "generally" I'd be prepared to make is that those being unpleasant about inherited wealth have different ways of doing it than those being unpleasant about non-inherited. However, of these differences, the old money unpleasantness tends to grate on me slightly, though not meaningfully, more.

Time Magazine has a separate version of its website with ads only AI can see

Re: Any lawyers reading this?

That is technically true. However, the two steps are being done by different people, which gives both of their respective lawyers ways to argue that they're not responsible in addition to the existing get out of consequences free fallbacks that apply to most LLM things:

Ad in paper lawyers: Everything we said there is valid advertising copy. It's not our responsibility if a chatbot that we didn't write or tell people to use isn't capable of generating good results based on it. Fallback: You have no proof the LLM told the user to use this bank because of our text; you can't track a prompt to a specific bit of training data like that.

Chatbot company lawyers: Our LLM is intended to search the internet and summarize what it says. It's not our fault if the pages online have text designed to show up in those summaries. We didn't write that or do anything specifically to promote it. Fallback: We wrote in the terms of service that this was for entertainment only and warned people to check the accuracy of results, so the user who used it for financial advice and didn't read any further was not using it as directed.

London cops handed victim's new address and number to her stalker, watchdog says

Re: Red Dwarf

It's a comedy show and the computer concerned is computer senile anyway but, if we want to get technical about it, that's not the right calculation either. While you can't just add IQs, you can't just average them either. That would, at best, get you the mean intelligence of a population. If they were working together, a bunch of people who have small abilities in different areas could combine their skills and achieve a higher score, so a group trying to complete an intelligence test might score higher. This is mostly dependent on their ability to recognize when someone's better at something and to let them do that part, not generally part of an IQ test.

'Asimov was right' about rules for robots, says ex-US Cyber Director

Re: zeroth law

I don't know, and I was not either of them, but maybe they disagree that that's the important factor to determining what someone's nationality was. He was born in Russia but left it for the US at the age of 3, never learned Russian, never returned to Russia (to live, I don't know if he traveled there), and didn't retain Russian/Soviet citizenship. I would not consider that a sufficient basis to consider his nationality anything other than American. I admit that there can be more to a nationality than the legal factors of citizenship and residency, but I think the other relevant and still smaller ones are culture and allegiance, not ancestry.

Re: Chris Inglis hasn't read Asimov

It's a reference to the current US Education Secretary who last year read out a speech written by someone else about AI use in schools without understanding what she was talking about, so she kept reading AI as A1.