Whilst it's not that easy to discover you shouldn't have to be holding shift before you start dragging, it's generally only whether you're holding it down when dropping the items that matters in KDE from what I've seen.
If I'm reading the article correctly, the headline claim is about a bug that was discovered in the original Horizon system right as it was about to be replaced with a new system, Horizon Online, at which point the bug fix would be completely useless. Other things like bugs in audit reports that continued to matter and affected the prosecution of subpostmasters were officially priorities, they just did a terrible job of handling them.
The same thing sort of happens with humans teaching humans though: you get things like linguistic drift where the structure of language and the meaning of words changes over time, and artistic styles and musicial fashions gradually shift as well. What's different here is that there's an ouside reality of what (for example) faces look like which human perception is grounded in but AI training doesn't have access to unless it's added to the training data. This means we also probably shouldn't worry about humans consuming AI generated stuff having the same effect, since there's no reason to think it should be any worse than humans consuming human-produced stuff.
Quite a lot of what news publications like the New York Times do is precisely regenerating articles that roughly approximate source material from some other publication. If I remember rightly, a lot of smaller, more local news organisations aren't happy about this because of course it's a more or less direct substitute for the original and a handful of big news organisations (particularly the New York Times) are taking so much of the money that people are willing to pay for news that it's affecting the viability of the rest - but it's not illegal, since facts cannot be copyrighted.
Wealth is indeed power rather than simply a collection of material goods, and most of all the wealth this article talks about is the power to run specific large businesses in the way those specific wealthy people want. That is, the way capitalism works in this case is that control over businesses and the ability to dictate how they allocate resources is given to individuals who are in some sense "successful" at running those businesses in that they manage to make lots of revenue and convince investors they'll do a good job of continuing to do so going forward.
The reasons why this system might be attractive may become more obvious if we consider some alternatives. For example, in Russia currently the power to control big businesses is given to politically well connected "oligarchs", and anyone whose business success exceeds their political influence tends to find that business is taken off them and given to someone much less capable but better connected. This does not lead to a functining economy. (There's a popular idea on the left that capitalism has oligarchs too which we just don't call that, which misses this important difference.) We could also imagine a system where this power is distributed democratically, and everyone in the country gets a say in how all the businesses are run. The problem with this is that in order to make decisions that are as well informed as even just some dude who happened to be in the right place at the right time, every member of the public would have to dedicate as much time to informing themselves as all of those random dudes collectively do already. That is, it's structurally impossible to just distribute this power out to everyone.
Also, most of the things that populist campaigners claim that wealth redistribution will achieve are about more material goods for everyone which relies on those material goods or at least the capacity to produce them being there already and just horded by the super-rich - just handing out power will not satisfy people's expectations for wealth redistribution and wealth taxes.
Yeah, Oxfam have been pushing a bunch of similar misleading claims using the same trick, such as arguing that people across the world lost vast amounts of pay during the pandemic even as the rich got richer by a vast amount and insinuating this was money that was somehow redistributed from one group to the other. Of course, this is just using the fact that share prices anticipate future changes in the economy to measure from close to the bottom of the pandemic-induced drop, which happened right around when the pandemic was first declared and it was just starting to affect people, and there is no actual mechanism by which money could be transferred in such a fashion. The money lost in pay was essentially destroyed; people didn't work, didn't produce things, and the real global economy of stuff available to purchase shrunk by if anything much more than that (due to furlough, unemployment, inputs other than labour, etc). The increase in wealth of the super-rich similarly was not transferred from anywhere, it was just the valuation of things they already owned going up due to the price shares were trading at increasing. This didn't mean an equal sum of money was somehow transferred into the stock market to do this.
Yeah, the big headline claim of Forbes' billionaires list from 2023 at https://www.forbes.com/billionaires/ is that half the billionaires on the list lost wealth over the previous year and that in total they were worth less than a year ago. I don't think that's adjusted for inflation either, unlike claims about workers' pay getting worse. There was also a drop off in billionaires' wealth in 2020, which coincidentally is where a lot of claims about the rich getting richer from organisations like Oxfam and publications like the Guardian start measuring from.
This article does not provide any actual evidence for that idea though. In particular, the richest person in the world - and I think possibly also the one whose wealth grew the most - wasn't even a billionaire until fairly recently, and the basic investment thesis that's made his companies so valuable is that they'll take a big share of the global car market, wrecking the income of existing car companies and destroying the wealth invested in them. Actually, I think all five of the people this article is about became billionaires in the last 30 years or so, displacing the previous richest people in the world as they did so.
You've also got to remember that by focusing on the richest people in the world right now the article is effectively cherry-picking the people who've done the best in terms of wealth whilst ignoring those whose wealth has failed to grow so well.
I do wonder how many distinct artistic styles Midjourney and other similar image generators actually have. Say that you you were to cluster all the styles it knows about based on similarity and then generate images of the same things with all of them - how far apart would the styles have to be before the difference between them is larger than the normal variability in output? It's my general understanding that getting a reasonably decent imitation of a particular artist's style is usually a lot more difficult than just putting their name in a prompt, that what that copies is generally fairly superficial and high-level stuff.
Last I heard, everyone was struggling to fine tune SDXL on furry art too. The only community fine tunes that seem to have succeeded are ones aimed at taking things it was already designed to generate and trained on and making it actually good at those, or improving the aesthetics of its output.
It's basically impossible for the developing world to copy the route to replacing coal usage that the developed world has taken though, because that relies heavily on natural gas and there simply isn't enough of it being produced, and of course the wealthier developed world gets first dibs. This has likely been made worse by environmental activists campaigning against new gas production. Countries like Pakistan have had to switch away from gas and launch emergency coal power building programs because they simply cannot keep the lights on otherwise.
Yeah, it's long been common for more traditional hosting providers to limit egress traffic and charge overage fees for going over that but have no similar limit on incoming traffic for basically the same reason: their network-wide traffic patterns mean that egress is what costs them money and ingress is effectively free on top of that.
ROCm is also incredibly fragile and buggy at the best of times, so anything not actively tested by them stands a good chance of not working. Hell, I remember that a while back people were having problems with machine learning code giving garbage results on one of the few consumer GPUs that was officially listed as supported and AMD eventually replying to the bug report and declaring that actually, no they weren't going to support it and they'd remove it from the list ratther than try and fix the issue. I think this was back when newer consumer GPUs were genuinely unsupported as in the code simply wasn't there too. Integrated GPUs have also alwayu had a lot of problems.
There's also questionable OS compatiblity. ROCm is Linux-based and has extremely limited and rather experimental Windows support. Their fancy new neural processing unit is Windows-only, tied in with a Microsoft framework, and they don't seem to have any kind of definite plan for supporting it elsewhere. So there's quite possibly no single OS where all the hardware that theoretically could be used for machine learning and AI in these chips actually has even vaguely functioning code that works on that OS to make it work that way.
If there's any policy which people are calling for but which the government doesn't currently support - for example, passing a law that blanket overturns all of these convictions - it tends to get portrayed as an obvious win with no downsides by much of the British press. If there's something which is currently government policy or effectively so, like leaving this to the courts, we hear about only the downsides of that policy. The BBC does this too, but not quite to the same extent, and they get a lot of flack for it as in this case.
Actually, the government has just announced that they are going to pass such a law, and this helps further demonstrate what I mean. After they announced this, the Guardian ran an analysis piece pointing out all the same reasons why this is potentially a bad idea that the BBC talked about before: https://www.theguardian.com/uk-news/2024/jan/10/no-precedent... Their top headline article is also about this announcement and leads by pointing out the same problems: https://www.theguardian.com/uk-news/2024/jan/10/rishi-sunak-... The way in which the BBC is more pro-Tory than publications like the Guardian is that they're less likely to ignore these problems when talking about activists calling for this prior to it becoming government policy. That is, in order for a British media publication to be adequately anti-Tory in the eyes of the press, whether they treat a proposal as bad or good has to depend on whether it's coming from the Tories.
Which is a distinction which basically doesn't matter for air purifiers like this one - it's only really relevant to stuff like whole-room positive or negative pressure systems which ensure all of the air goes through the filter. With a standalone air purifier, the main limiting factor is how quickly the room air gets exchanged through the filter. It's pretty much irrelevant whether the filter removes 99.95% of particles or merely 99.5% since that's outweighed by all the pollution in air that hasn't been through the filter at all yet. The claims about why true HEPA filters are better are nonsense.
The Met's had an investigation open over this since 2020, all that seems to have happened as a result of this popular four-part series is that they've confirmed they're looking into specific possible criminal offences by the Post Office or its employees that we didn't know they were looking in to before. It's not clear that this is something which started investigating because of the popular TV series, but most likely not and they were just pressured into revealing more about the existing investigation - the police here don't generally give running updates on criminal investigations to the media.
One of the big problems with British politics at the moment is that lots of people seem to think that if the BBC is less willing to pretend there are easy solutions with no downsides the government is ignoring than other parts of the media this is seen as them being in the government's pocket. Which isn't the same as not willing at all, they pretty systematically focus only on the downsides of government policies already - one of the examples that made HN was the UK Covid app, where the BBC treated using Google and Apple's framework as obviously better right up until the point the government launched a new app that did, when they immediately switched and focused only on the advantages of the old app due to it not doing so. They just don't do this as consistently as some other outlets.
That is, the minimum standard they need to meet in order not to be perceived as carrying water for the government is not just lying and misleading the public in order to get the ruling party kicked out, but doing so as aggressively and shamelessly as the most aggressive and shameless of their competitors (which includes social media to some extent).
I don't think that poses the same problems for judicial independence because the courts were correctly applying the law at the time - it's just that we've now decided that law was morally wrong - and of course it's always been Parliament's job to set the laws. However, in this case there's no dispute that defraudingg the Post Office out of large sums of money should be a crime, the problem is the courts' factual findings that lead them to rule that specific people had done so.
While the LAION data set is sloppy in a whole bunch of ways, this is also just a mess in general. The researchers found apparent CSAM hosted on a bunch of mainstream sites - Reddit, Twitter, Blogspot, Wordpress, XHamster and XVideos. Also apparently a whole bunch of "teen models", nudism, and Japanese "junior idol" content which... well, let's just say that a lot more things used to be legal in the relatively recent past that probably shouldn't have been and leave it at that.
As far as I could tell, the policy allowing artistic nudity never existed in practice and Twitch doesn't seem to have been serious about it. Even though it was supposedly allowed, basically everyone who drew art with genitals or female nipples got their account suspended aside from one or two exceptions that seem to have slipped through the cracks. This wasn't just limited to AI art or anything even vaguely photorealistic either, and it included Twitch partners whose bans are supposedly are handled in house. Also, whilst the Twitch CEO did say that some people had been accidentally banned who were following the rules, Twitch conspicously didn't unban any of them until comfortably after the U-turn. There was also nothing stopping the full, uncensored thumbnails showing up on the Art category page, including for users whose accounts were marked as under-18 - which meant that it was full of naked anime titties.
Yeah, but if the media did that rather than cherry-picking then it wouldn't be possible to create an endless stream of headlines like this which make it seem like the richest people always win and this must be somehow at the expense of the rest of us. I remember there were a bunch of headlines about the wealth of the richest increasing during the pandemic that pulled the same trick too, measuring from around the bottom of the huge pandemic-induced dip in the markets.
I think there's another layer of cherry-picking too, in that the list is the 25 richest families in the world now, so it excludes families whose investments have done poorly. Not sure how much of an effect that has, but entry number one and number five weren't on the list last year and they're both owners of big state oil companies which is a sector that's done unusually well compared to the rest of the economy and to how it was performing a couple of years ago.
The big practical obstacle to this is that the things this flags people as potentially committing fraud based on look to be the same things an actual fraudster would manipulate in order to get payments they're not actually entitled to - stuff line income, living situation, kids etc. So the obvious route to not setting off fraud alarms would likely involve not committing fraud, or at least not the most obvious and profitable kinds.
Yeah, I'm sure the performance of this must be pretty bad. As I understand it, non-batched LLM inference is generally limited by the bandwidth required to read the entire weights from GPU memory for each token produced. This trick instead moves the bandwidth bottleneck to loading those weights into the GPU from wherever has enough space to store them, probably some kind of SSD, and that has much lower bandwidth than GPU RAM.
As the original article explains, whilst the inactivated polio vaccine used in developed countries like the USA does provide good protection against being paralyzed by polio it's not terribly effective at preventing transmission. So it's not actually all that clear whether Rockland County's attitude to vaccination will impact anything other than the people living there in this case.
The server-side Matrix code developed by Element is no longer going to be released under the Apache 2 license. It's going to be dual-licensed AGPL/proprietary from now on and released from Element's own GitHub repositories with a CLA requiring copyright assignment to Element so they can use it in the proprietary version, and the foundation is dropping all development on the Apache 2 version since they're not interested in competing with Element. This was specifically to obstruct other companies from providing Matrix-based enterprise offerings. There was an announcement about it which made it onto HN a while back: https://news.ycombinator.com/item?id=38162275
This reply about how in reality there's very little benefit to the community from the copyright being assigned to the foundation is a lot more perceptive than that comment, especially now that exactly what it foresaw has come to pass: https://github.com/matrix-org/matrix-spec/issues/571#issueco...
That's a clause from the standard GPLv3 covering something different:
"An interactive user interface displays “Appropriate Legal Notices” to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."
The requirement to offer users of modified versions interacting remotely over a computer network a copy of the source is seperate and has no such limitation. Which rather suggests that the drafters of the license considered this issue of requiring people to make unrelated modifications and did nothing to stop it.
The AGPLv3 is designed to ban versions with proprietary alternations full stop. In order to do that, it places some pretty daunting requirements on any modification - even open source ones. The moment you make any change, you must ensure that all users interacting with the software over the network are offered an opportunity to receive the corresponding source and that offer must be prominent. The original software does not have to include a way to offer such notices, just updating any existing code that does to point to the modified source is not necessarily sufficient to comply (unlike with the previous Affero GPL), and the protocols it relies on don't even have to be designed in a way that makes this feasible. It's not clear how this would even work with a federated chat protocol.
Note that one of the projects on this list - Element - recently changed their license to one intended to make it impractical to deploy the open source version and make users buy commercial proprietary licences instead for the specific reason that governments were buying support for their software from companies other than them and they wanted to stop that. Which I think shows a pretty big practical limitation of this approach.
From what I can tell the star that it occulted in September was much too dim to be visible with the naked eye, whereas this is one of the brightest and most easily visible and distinctive stars in the night sky.