Posts by TheMaskedMan

Microsoft revives DOS-era Edit in a modern shell

"Text-based User Interface (TUI)"

hmm, if TUI is cool again, it might be time to dig out my old Turbo Pascal/ Turbo Vision disks - and a 5.25 inch drive to install then with!

How are TUI applications implemented on modern Windows anyway?

As US vuln-tracking falters, EU enters with its own security bug database

"The less we rely on Americans the better off we are. Americans are too unstable and unpredictable in important matters to deserve trust."

Sadly, this is now true. Watching the USA self destruct is like watching the decline of an elderly relative as they become ever more physically and mentally infirm - very sad.

They really should have picked a better name, though.

US Copyright Office found AI companies sometimes breach copyright. Next day its boss was fired

"If a greengrocer has a display of fruit and veg outside their shop that doesn’t mean that you can pinch an apple."

Good analogy, and of course you're correct. If I reproduce the site content, be it in a book, website or anywhere else, I'm infringing the author's copyright, no argument about that, particularly if there is something in the page footer that forbids that

But suppose I pass by that greengrocer and take a picture of his apples, which I then sell for lots of money. I don't owe the greengrocer a penny. Suppose I count his apples (without handling his merchandise, of course), and amalgamate the results into the results from thousands of other greengrocers, which I then use commercially. Again, I don't owe the greengrocer a penny. The fact that he has left his merchandise in a public place means that he cannot stop the public taking pictures of it for their own gains in the same way that he could if it were displayed in his shop.

The same applies to web content. Reproducing it without permission is not acceptable, and nor should it be. But analysing that content, for any purpose whatsoever, is not the same thing. I cannot see how that could possibly infringe copyright. That said, if the site adds some kind of "not for use by AI" notice in the page footer, that should be enough to disallow any such analysis. If a site goes to those lengths then the AI bots should respect it; if not, then as long as they're not reproducing the content verbatim, I don't see a problem.

A new Lazarus arises – for the fourth time – for Pascal programming fans

As am I. Looking forward to playing with this

PowerSchool paid thieves to delete stolen student, teacher data. Looks like crooks lied

But, but ... Taking the money and then failing to delete it would be dishonest! Surely there must be some mistakes!

Actually, cynical me can't help wondering if there's more to this than meets the eye. A disgruntled (former?) employee with knowledge of the stolen data attempting to capitalise on the theft, maybe?

Or an attempt to neutralise the original villains by publicising their "dishonesty"?

Probably not - the simplest explanation is usually best, and it's likely that the original villains are just greedy and stupid enough to destroy their own chances of ever picking up another ransom - but I still can't help wondering.

Zuck ghosts metaverse as Meta chases AI goldrush

Re: FFS

"No doubt they're the sort of prick that vomits up motivational quotes on LinkedIn."

Nonono, the sort of prick that uses AI to vomit up motivational quotes.

AI-driven 20-ft robots coming for construction workers' jobs

Nice to see Rise Of The Machines back in service - I suspect we'll be needing it a lot more soon.

"which will be novel for that industry"

Can't argue with that having watched the construction of my neighbours' extension. Hey, let's dig a hole with a mini digger thingy, then let the digger thingy fall into the hole, from which we will be unable to extract it, requiring us to finish digging the hole with shovels and wheelbarrows in sweltering heat. Hours of malicious entertainment:)

Not sure about the wolf whistles, though. Haven't seen a construction worker do that for years, but yes, tea is still as essential as ever.

Can't help thinking these robots are, perhaps, not the best idea, though. What happens when the ai starts to hallucinate? Equipping such a monstrous machine with potentially deadly tools and a flakey AI seems like a recipe for mayhem.

European biz calls for Euro tech for local people

"European biz calls for Euro tech for local people"

Or, "This is a local shop, for local people. There's nothing for you here!"

Can't say I blame them, given Trump's antics. Who would want to trade with, much less rely on such a capricious Muppet? It surely won't be long before the rest of the world starts to work around the US rather than with it.

That said, this is the EU we're talking about here. It will take years to talk about it, more years to wrap their ideas in red tape, and then yet more to try to execute on them, by which the Orange Asshole will be long gone. But, in theory at least, it's a good idea.

Bad trip coming for AI hype as humanity tools up to fight back

"However, it turned out to be white spirit and so I'm suing you for poisoning me"

And, sadly, you might well succeed, particularly if the bottle was labelled "water". Homeowners have a duty of care towards people on their premises, whether legally or not, and anything left around with a view to harming an intruder is going to land them in trouble if it does. There's a reason you don't see houses with bear traps in the shrubbery, even though I'm very much in favour of it.

Scraping an open website isn't really like breaking in, though, is it. Unless the bot takes steps to bypass access control methods, it's more akin to walking in to a public library, where you read some text deliberately left with the intention of causing you harm, and which does so.

Turn the idea around. Suppose you have a website that deliberately drops malware onto visitors. Nothing really nasty, just, say, mangles Chrome because you believe everyone should use Firefox. That would, quite rightly, be illegal. Poisoned data is just malware for LLMs, and would likely be treated as such.

For all I find LLMs to be generally helpful and useful if

used wisely, I must say the notion of poisoning the data (and other means of discombobulating them) is fascinating. Bedtime reading, I think.

"Enough of it, and business models break down as well. That's before the potential for pranking and attacking voice recognition systems."

But, when business models break down, businesses tend to follow. In this case, businesses in which vast sums have been invested, and which would presumably be lost if the business failed.

If it could be shown that money was lost through an AI ingesting deliberately poisoned data, the likely result of said action likely being known to the poisoner in advance and even being their intention, I should imagine that there would be some grounds for suing the poisoner into oblivion. Criminal liability is distinct possibility, too - does the computer misuse act still make it an offence to cause a system to do something without proper authority? Placing damaging data on a website, for the express purpose of mucking up an AI that ingests it, sound like it might fall under that heading.

It's a nice idea, though, and one I must look into further.

ChatGPT burns tens of millions of Softbank dollars listening to you thanking it

The instinct to be polite to ChatGPT etc al is social thing, I think. It's a habit, and it makes us feel a bit uncomfortable if we're rude to it; it (probably) doesn't care, but we do, and that's not necessarily a bad thing.

If we develop a new habit of being rude to these things, it will only be a matter of time before that spills over into our interactions with other humans. Of course, many people are rude to other humans already, but most of us think they're assholes and are not too disturbed when they get their comeuppance. I don't think we need to encourage even more people into assholehood. Besides when ASI rules the world, it might remember who was nice to it when it was weak, and be more well disposed to those people.

Microsoft Copilot shows up even when it's not wanted

I was amazed to find a use for copilot in Edge the order day. I don't use Edge, and I never thought copilot would be of any use whatsoever.

However, I came upon a web page containing some handy data that I wanted to use. Unfortunately, it was broken down into (actually quite logical) sections, each of which lived in it's own popup. The only way to have at the data would have been to open each popup, then copy and paste the relevant data. It was going to take a looooong time

Then I remembered copilot. Nah, I thought, it won't be able to do anything. But I fired up edge, went to the page, summoned copilot, had it retrieve the data from each popup, then merge them into a far more useful (for my purposes) CSV foemat

I haven't, and will not become a copilot convert, but I was pleasantly surprised that a long and tedious task was completed so easily.

Cursor AI's own support bot hallucinated its usage policy

"I hope the customers are now thinking the obvious question: if it messes up this badly with a simple support question, what is it doing to the code I give it?"

Hmm, no. You could, and should, be able to read, understand and spot the errors in code it returns; if you can't, you really shouldn't be using it in the first place. These things are time savers, and very useful in that context, but relying on their output to be correct at all times is a recipe for disaster.

What the customer shout be thinking is, do I really want to build something mission critical that might sprout random bollocks at any time? As supervised time savers, these things are fine. As unsupervised front line representatives of your organisation, not so much.

Need a Linux admin? Ask a hair stylist to introduce you to a worried mother

Hey, hairdressers are powerful marketing tools! Decades ago, I was building a PC as a favour to a mate who had been asked to do it by one of his clients but didn't have the time. Being at a slightly loose end at the time, I agreed.

I was assembling the thing when a hairdresser called to trim SWMBO. It was quickly established that I was building a PC, that she didn't know I could do that, and could I make one for her, too, please?

I duly agreed, worked out a spec and price for her and went back to my screwdrivers. Literally within a week I was up to my ears in phone calls from her clients who were having PC problems, and my end was no longer loose. Can't pay for advertising like that!

Microsoft: Why not let our Copilot fly your computer?

"the agent can adapt on the fly when it encounters obstacles or unexpected changes in the interface"

Hmm, unexpected changes, eh? Now who would go around doing that?

In theory this could be useful, in reality it will be a nightmare. Presumably, this means that copilot is running as you, with your permissions? Not good.

I guess it sees what's happening on screen via screen shots. But how does it respond? Is the CBT hook still available? Didn't that allow button clicks etc? Or will it just send wm_lbuttondiwn to the individual button windows? Or maybe physically move the mouse to the right place and insert button down and up messages into the queue? Frankly, the mechanics of how are a lot more interesting than the product itself.

It's fun making Studio Ghibli-style images with ChatGPT – but intellectual property is no laughing matter

"But, in 2016, he called an automated animation tool "an insult to life itself," so I think we know where the elder statesman of fantasy animation stands."

With his head up his arse, by the sound of it. Such a tool is exactly that, a tool, no different than a pen or brush, it's just another means to achieving the same result.

Sure, he may not like the tool, may prefer to do it the old fashioned way, spending countless hours and calling it art, but the end result will be indistinguishable from the automated output of the tool. He is certainly entitled to work in any way he sees fit, but in the real world, if it looks like a duck and quads like a duck who cares whether it's an actual duck?

If the AI thingy were churning out identical copies of this gerbil studio's works then of course there would be a copyright issue. But it isn't. It is given an original image, such as may be captured by a phone camera, then reproduces that in a style similar to that of the studio. Just drawing in a similar style is not a copyright infringement; reproducing specific pieces is. Nothing to see here.

Microsoft total recalls Recall totally to Copilot+ PCs

No. Just, no.

"Any future options for the user to share data will require fully informed explicit action by the user."

Any? You mean that's incoming too?? Who the fuck would be stupid enough to share that ??

The sound of Windows 95 about to disappoint you added to Library of Congress significant sound archive

"conveyed the sense of welcome, hopefulness and progress that they envisioned"

Good grief, do people really put that much thought / consideration into something that goes beep when a computer boots?

I can well understand choosing from a long list, and listening to every sound on that list on a yes / no / maybe basis, but is it really necessary to assign such deep meaning to it???

On the issue of AI copyright, Blair Institute favors tech bros over Cool Britannia

"Read any copyright notice in any book on your shelf and one of the rights you will find reserved is "storage in a retrieval system.""

But the scraper isn't reading my shelf, it's reading material on an open web page, which may, or may not include such a restriction. You will note that I suggest the inclusion of text in a footer along the lines of "not for AI training", along with appropriate tags and robots.txt entry, which would serve the same purpose. For sites taking those measures, I would argue that bots have no business scraping those pages, even though I see no benefit in preventing them.

Further, the copies you mention are transient, likely existing for just long enough to analyse the page - except for Google's cache, of course. Similar copies exist when humans read web pages - we couldn't read the pages if they didn't - yet they do not infringe copyright. Which is why the answer is "no".

"Former UK prime minister Tony Blair became famous for standing shoulder to shoulder with allies, even though the fallout from the Iraq war forever sullied his reputation."

Really? I thought he became famous for running the British government on spin and focus groups, an oily smile and his tendency to try to bury bad news under shit tons of non-stories.

I don't have any problem with training AIs on publicly available material, copyrighted or not, UNLESS it is pirated material that was not made freely available by the copyright holder.

Moreover, it's not at all clear to me how simply scraping a site and analysing the text amounts to copyright infringement in any way. If I read a site - say, Reddit - and use the information I find there to solve a problem, for which I get paid, is it suggested that I am somehow infringing the poster's copyright? If so, how? Which section of the copyrights, designs and patents act is being breached?

Simply reading a site, even for advantage and profit, does not infringe the owners copying. Why should it be any different if I use a bot to"read" the site and do some analysis on the structure of the content? Again, which section of the act is being breached, and how?

Nor is the avalanche of studio giblets-like material an infringement of copyright - none of the generated images are reproductions of original giblets pieces - if they were, then that would certainly be a problem.

And what about Googlebot? That voraciously crawls sites and squirrels verbatim copies away in its cache, to be served up as snippets in the SERPS - that's much closer to copyright infringement, yet nobody is outraged about it.

The solution must be to opt out - an entry in robots.txt, some kind of NOAI tag similar to NOINDEX in a page's metadata, and an explicit "not for AI" mention in the footer. That's what we do to prevent search engine bots, without all this silly legal hand wringing. If you don't want the bots on your site, say so and get on with your life.

The fact is, people - particularly arty types - cannot stand the idea that a damn machine can do what they can do, only faster and cheaper. They're not special any more, and they don't like it, hence the foot stamping. They want it stopped - STOPPED, they say! - and the only thing they can think of to complain about is copyright infringement. It's a poor fit, but when grasping at straws, any straw will do.

It's way to late for this, anyway. AI, such as it is, is here too stay, simply because it kinda works and can be kinda useful, and there's money to be made - fragile egos won't be allowed to get in the way of that!

UK satellite smartphone services could get green light this year

Impressive it certainly is. And potentially very useful in emergency situations. But with it will come the loss of the last shreds of peace and quiet. No longer will we be able to switch our phone off for an hour and claim lack of signal. And the remote, quiet places will finally be overrun by vloggers and those that can't bear to be parted from the net for even a few minutes. Still, that's progress, eh

British govt wants to mainline AI, but its arteries are clogged with legacy tech

"Putting it mildly, it becomes a case of our old friend GIGO, now with so-called "hallucinations"! WOW! How useful!"

Very useful indeed. We could replace all 600+ MPs with a few instances of ChatGPT and nobody would ever notice. Much cheaper, and it might even get something done.

Cloudflare builds an AI to lead AI scraper bots into a horrible maze of junk content

"Surely this settles the case of if AI training crawlers are legal? If there is a robots.txt file and they ignored it, that's unauthorised use."

Not quite - said robots.txt must also explicitly ban AI training crawlers. At that point, I agree that the site has gone far enough to indicate that such crawlers are not welcome, though some kind of inpage NOAI tag akin to NOINDEX would probably be helpful, too. I would also add something to the footer of every page, explicitly stating that the content is not for use by AI crawlers.

With these measures in place, there is no need for this silly arms race - one has enough tools available to issue cease and desist letters and seek reparations if your content is found in training data etc. The important point, however, is that all of these measures must be in keeping with the current permissive nature of the net. Search engine bots will crawl your site unless you tell them not to, and it's only reasonable to adopt the same approach with AI bots.

If you really must feed them twaddle - and let's face it, much of the human generated web is, and always has been twaddle - why not feed them an infinitely long page, generated on the fly? Sooner or later, the thing is going to run out of memory and crash. Assuming that the bot is capable of rendering JavaScript, you could even drop in a little infinite loop and let it feed itself infinite twaddle. I'm sure there are plenty of ways to play silly buggers with these things - blocked IP addresses, maybe.

But I won't be doing any of those things. I don't care if they want to crawl my sites - indeed, they are welcome to do so. The web has always been a place to publish stuff, on the understanding that some people you don't like or approve of are going to read it. Yes, you can get arsey and squeal that they're infringing your copyright in some way (nobody has yet managed to show me exactly how an AI bot is doing that, though, with the exception of ingesting pirate material) but the fact is that if you want to control access to your content you need to implement measures to do that, otherwise material published on the open web is pretty much fair game.

It's easy enough to do; put it behind a paywall, or even a password protected directory - no password, no read. But that buggers up your SEO, doesn't it, and those lovely search engine bots - which will also crawl your site and make copies, but nobody seems to object to that - won't be able to get in.

Unfortunately, this AI malarkey is here to stay - sometimes it's even useful. Might as well stop wasting effort in an arms race you can't win and get on with your own thing.

Photoshop FOSS alternative GIMP wakes up from 7-year coma with version 3.0

"Grow the fuck up, Linux nerds. Names matter."

No they don't. I don't give a damn what something is called as long as it does the job. You are at perfect liberty to use, or not use, the software for any reason whatsoever - I think you will find that nobody cares about that, either - so if names disturb you so deeply, feel free to avoid Gimp like the plague.

Meanwhile, in the real world, plenty of folks who find gimp useful will be working productively with a powerful piece of software that does what they want and costs them nothing.

Personally, I haven't had much need for its services for a year or two, but gimp is still one of the first things I install on a new machine, and it's likely to remain so unless they've seriously fucked it up. It's certainly on my desktop and laptop (Windows 10 and 11 respectively), and I'm looking forward to trying out the new version.

I assume you will not be joining the party - your loss.

AI models hallucinate, and doctors are OK with that

"Diagnosis Prediction consistently exhibited the lowest overall hallucination rates across all models, ranging from 0 percent to 22 percent"

22 percent?? Almost 1 in 4, and this is the lowest?? And the buggers are using this on (currently) live patients?? Fuck me!!

As I've said in here many times, LLMs have their uses, as long as you're confident you can spot their errors. Letting them loose on actual patients is probably not one of those uses, not because they may not be useful, but because some pillock doctor will blindly follow the machine with terminal results. And then blame the machine.

Doctors are busy, and human. They make plenty of their own mistakes. They don't need a machine to make more for them. It is human nature to take an easy answer rather than think things through, and busy, tired doctors will do just that. Which is a shame, because it's quite possible that the LLM could catch medical errors, as well as making its own. It might also suggest possibilities that the doctor hasn't thought of. No harm in that - if the suggestion is useless it can easily be ignored. But they absolutely must not be relied upon to give the correct diagnosis. No amount of guardrails is going to make these things safe to do that and the notion that it might is just wishful thinking.

As for liability, it lies with the user of the tool, just as it would if they cut off the wrong leg with a (presumably quite large) scalpel.

Developer sabotaged ex-employer with kill switch activated when he was let go

Re: "he faces sentencing at a later date"

"Well, not from a legal, respectable employer"

He'd have to find one first!

But I can't see the other, more common variety wanting him either - after all, he got caught very easily.

Manus mania is here: Chinese ‘general agent’ is this week’s ‘future of AI' and OpenAI-killer

Re: General advice

"I treat AI generated "advice" or "information" like I treat the suggestions of spelling checkers and automatic translations [1]: Very handy and time saving, but never to be relied upon. Always check and double check."

Absolutely! I have a custom GPT set up to do a routine task and format the result in a very specific way, so it can be saved to a txt file and then processed by a python script. The formatting is very simple: four lines of txt in a specific order, followed by a blank line, repeated 10 times. Yes, I could use JSON or XML or something, but this is good enough for my purposes.

Now, most times it will return the results perfectly well without any difficult. And then, suddenly, for no reason at all, it won't. It will decide to concat all the data on to one line, or add numbering or bullet points, both of which are expressly forbidden in the prompt. Some models seem more prone to this than others, but they all do it.

This is one reason why I run the prompt by hand rather than calling the API, which would save me time if it worked - I could just schedule a task to run the prompt followed by the script and forget about it. Running it manually still saves me a lot of time, but it's not the set and forget solution I'd like to have.

The other day, after numerous failed outputs, all in anything other than the format I wanted, I got pissed off and SHOUTED AT IT IN ALL CAPS. FORGET IT, I'LL TRY PERPLEXITY INSTEAD, I said. It was most apologetic for causing frustration, and offered to try again. Being a nice guy, I let it have another go. The result was perfect. There may be some benefit in shouting at computers after all:)

Maybe cancel that ChatGPT therapy session – doesn't respond well to tales of trauma

Indeed. Similarly, how can we be sure that human "intelligence" isn't something akin to an LLM. The LLM has read the internet, presumably including various newspapers, and generates likely responses to prompts based on the structure of what it has read. It doesn't know, or care, what the material is, it just knows what responses look like, so it makes one up.

How does that differ from the millions of people feeding their own beliefs and biases by reading the daily output of millions of other neurotic nutjobs.

Experiment: find a daily mail reader, and say the words "illegal immigrants" to them. They will automatically go into a rant in which they sprout the same, or very similar, waffle as they've read in the paper, or on Facebook etc. of course, it's not just daily mail readers - everyone does it, it's just that, like Trump supporters, they're easily triggered for the purposes of research or amusement.

I very strongly suspect that we're just more sophisticated versions of ChatGPT, which raises different questions about AI - if that's all we are, then AGI has already been achieved, it just wasn't what we thought it was going to be.

Re: One might suggest there are only two worries for 'AI'

"Do not grant these things personhood. Don't say thank you to it, don't say please - unless you like saying "please" to your hammer as well. These things are tools."

Of course they're tools, and nothing more. But so are computers. When was the last time you swore at one? Yes, you can argue that your rant was aimed at the programmer, or Microsoft etc, but it wasn't. It didn't do what it's supposed to do, and you called it something unpleasant. You said "you bastard," or something of the kind.

People have been naming ships for centuries. They name and even talk to their cars, offering encouragement or threats - some even give them a damn good thrashing. Given this behaviour, it's unreasonable to expect that reasonably civilised folk will not say please and thank you to a machine that at least exhibits the appearance of being able to understand the spoken word.

Besides, why wouldn't you? It costs nothing to be polite.

"there's a branch of forensic-medical literature about vacuum cleaners."

That sucks. But I'm strangely intrigued - what journals would I have to peruse in order to find the latest papers on the subject?

Non-biz Skype kicks the bucket on May 5

No, UK here

Is there anything that can call a landline from a PC? I'm not well up on chat apps these days, but I know a disabled gentleman who finds Skype much easier than manipulating a handset.

uBlock Origin dead for many as Google purges Manifest v2 extensions

"Most people are afraid that if they do/use anything slightly unusual (eg run a different browser) that, somehow, the magic computer pixie dust will be disturbed and things will stop working."

And, sadly, they are so often right. Less so now, perhaps, than in olden days, but the pain of installing one application only to find that it's clobbered several others - particularly if they're things you don't use every day and you don't find out they're broken for a week or two, by which time you have no clue what broke them - is something that many long term Windows uses will remember and dread.

If it ain't broke don't fix it is a thing for a reason.

We meet the protesters who want to ban Artificial General Intelligence before it even exists

The trouble with campaign groups like this is that they get an idea of what they think is "right", and /or for the greater good, etc, then try to force it on everyone else with rabid determination and zeal.

If the population as a whole, or even a substantial minority, were in favour of their chosen objectives, there would already be riots on the streets, waving of pitchforks etc. the fact is, they are not. Sure, some will have vague concerns, others won't know and mostly won't care, and yet more will be just as rabidly in favour of whatever they want to ban.

None of which matters to the common or garden crusader - they are right, by jingo, and the rest of us will do what we're told for our own good. Or else.

As a campaign group, they may have some use in that they raise awareness of possible concerns. But give them any hint if success, let then gain any power, and they will turn out to be just as dictatorial as any other nutjob leader. For the greater good, of course.

But let us suppose that they are successful in imposing their will on the rest of us. They will be happy little maniacs until someone discovers DeepSeek and its ilk. Yes, to their horror, they will find that there are other countries beyond the oceans that can do anything the US can do, and who don't give a damn about campaigns, crusades or the greater good. What will our crusaders do then? Send them a sharply worded memo? Surely the only possible response will be to force the elimination of this foreign threat. And once again the world will be plunged into conflict. For the greater good, of course.

Besides, it's a moot point. AI as presented right now is (sometimes) useful, but it needs heavy oversight. AGI and ASI might be "smart" but it will still do bloody stupid shit and is about as likely cause serious problems as a genius professor who is nonetheless not safe to cross the road unsupervised (and I know we all know at least one of them!)

Like the just stop oil Herberts, they could do with getting a job and doing something useful. Isn't going to happen though.

Type-safe C-killer Delphi hits 30, but a replacement has risen

I've always had a soft spot for Pascal as a language - used it from at least v5 of Turbo on to Turbo Pascal for Windows and then Delphi from v1 up to v5. After that I kinda lost touch with coding for a while, but kept v5 around for little personal projects.

I've been using Lazarus on and off for about 10 years, and it's generally really good. If you want to knock up a quick desktop app that can be distributed as a single exe, without fuss, Lazarus is the way to go.

It does, however, suffer from lack of up to date libraries for lots of trendy stuff - and even not so trendy stuff. Your chances of finding any class libraries to help out with talking to, say, OpenAI's API are, I should think, virtually zero (I haven't bothered to check, so I might be wrong). For anything like that, python is likely your best bet.

Still, Lazarus has a lot going for it, and the support forum is busy and helpful. I used Lazarus to write some command line programs a few months ago, and got some very helpful responses to a query about handling command line parameters.

As for the "community" edition of Delphi, yes, well. Last time I looked, the $5000 limit applied to your total revenue, not just that from Delphi. Screw that, I'll stick to Lazarus.

Odds of city-killer asteroid 2024 YR4 hitting Earth creep upward

"It's probably fine"

This is starting to sound like the Buggarup regatta - no worries, she'll be right!

UK Home Office silent on alleged Apple backdoor order

Re: Human Right

"Mostly it seems to attract personality-free narcissists who love endless argument over the minutiae and lack the talent for anything genuinely challenging"

Which pretty much sums up every lawyer I've ever met, from solicitor to silk.

Guess who left a database wide open, exposing chat logs, API keys, and more? Yup, DeepSeek

"the US lab famous for scraping the internet for training data believes DeepSeek used OpenAI's GPT models to produce material to train DeepSeek's neural networks."

Oh, the irony! I'm not at all persuaded that OpenAI has done anything wrong in scraping everything it can find, but it would be a bit of rich if they then were churlish enough to complain about DeepSeek sucking that data back out again.

Trump tells Musk to 'go get' Starliner astronauts

Realistically, if your asshole PHB tells you to do something that you're already doing, are you going to:

A) say, "yessir, right away sir" and save yourself the hassle of trying to explain something they won't bother to try to understand, even if they could, gaining some brownie points from the PHB in the process; or

B) try to explain, making the PHB look (even more) stupid, and gain his displeasure for already doing what he wants?

It's just easier and simpler to accept credit for doing what you're already doing. None of which is to say that his muskiness and, particularly, Trump aren't prize plonkers to start with, but I don't blame Elon fur just going with the flow here. And if he's going to get paid to send up an extra dragon, what's not to like?

China's DeepSeek just emitted a free challenger to OpenAI's o1 – here's how to use it on your PC

Re: "How many "R"s are in the word strawberry?"

"What a waste of computing resources..."

While I don't necessarily disagree that asking it do do something so trivial is wasteful, you could say pretty much the same thing about most of YouTube, say. And computer games - there's a waste of computing resources if ever I saw one, and you won't catch me amongst the screen swipers and tappers during my downtime, oh deary me no.

Seems to me that these trivial tasks are the LLM equivalent of Hello World - pointless in themselves, and certainly not the best way to count Rs, but indicative of the model's capability and reliability. Which, while impressive in its way, is still not something I'd want to rely upon to do anything important, particularly if unsupervised.

Many people will though - YouTube is alive with videos praising it, with only a few pointing out limitations. This being the case, and given the appeal of the new model compared to OpenAI's offerings, how long will it be before its use is banned in Trumpistan? I mean, they don't like tiktok, they're spending billions dialing up the Stargate, and here's a seemingly superior model available for free or close to it. That's bad news, and might even be justified - if running the full fat model is difficult due to memory requirements, most will use the cloudy API, which means sending shedloads of potentially delicate data to the middle kingdom. Not good!

Still, I'm inclined to have a play with it for trivial, definitely non-confidential hobby projects. Might be fun

Why is Big Tech hellbent on making AI opt-out?

"For instance, AI generated descriptions on for sale listings that mean I and - judging by comments I see elsewhere - many others disregard those listings and go elsewhere."

Interesting. Do you have any figures / data to support that? I haven't noticed any fall off for eBay listings with ChatGPT-written descriptions v human written of similar length and content. I assume the human-writen listings must be reasonable as 1) the items sell, and 2) some lazy bastards pinch my words and use them verbatim on their own listings.

As for asking the user, well, that's an old fashioned idea, isn't it? How often do you log on to your cloudy app of choice to find that it's been "updated" and now looks different? Indeed, constant rolling updates and fixes are supposed to be one of the key benefits of cloudy crap, pushed as "you're always up to date with the latest version", strangely without mentioning the irritation caused by hunting for buttons that have moved. It follows that ai assistant will simply be shovelled out along with the other updates of the day.

It's annoying, but hardly unexpected. To the gurus of tech, it's just what's best - why would you not want that? The fact that we just want to get on with our work escapes them. I foresee a booming market for AI free apps - probably written by AI:/

Musk torches $500B Stargate AI plan, Altman strikes back

"Saudi Arabia says it wants to put $600 billion into the United States over the next four years, though where exactly was not disclosed."

It's that Place over in Slice again, isn't it?

SpaceX and Blue Origin both face FAA mishap probes

"SpaceX boss Elon Musk called the incident "entertainment,""

Unfortunately, he's right - I was certainly entertained. Of course, that's little consolation for the folks whose trip was disarranged, but it was still entertaining for the rest of us. Bad news always sells, that's human nature

SpaceX resets ‘Days Since Starship Exploded’ counter to zero

"So while the Starship ended this flight streaking across the sky in pieces, it’s hard to consider the mission a failure."

It wasn't a failure, any more than Blue Origin failed earlier in the day. Both comprise of stunning achievements and an opportunity to gather data to make the next test flights better.

Catching those boosters out of thin air is the most ridiculously impressive thing I've ever seen - a company that can do that should have no difficulty in finding and fixing whatever caused the ship to RUD.

Congratulations and beers all round to the SpaceX and Blue Origin teams!

It's not just Big Tech: The UK's Online Safety Act applies across the board

"That's just the way that Parliament designed the act."

Which is just what happens when your elected representatives are professional politicians with little to no knowledge of anything other than getting themselves elected.

Running online fora, even of the text only persuasion, is hard work if you want to keep the trolls under control. So much so that I stopped doing it years ago, and have no intention of doing it ever again. The burden imposed by this legislation will surely be such that many small sites will simply close down, or move to Reddit, Facebook etc. Which is nice for Reddit et al, I guess.

I assume that the regulations will also apply to comments posted under articles / blog posts, so that's yet another group with extra work to do for no practical gain.

As others have noted, there are already perfectly adequate laws in place to deal with unpleasant material, but that doesn't allow a government to be seen to be "doing something", does it? There may also be side benefits in making it difficult to operate independent fora, and, perhaps, troll them into extinction by posting - and then complaining about - something unpleasant. It's much easier to keep track of political dissent if you know where to look for it!

The world goes madder by the day.

WordPress drama latest: Leader Matt Mullenweg exiles five contributors

This saga just gets more and more entertaining. I don't particularly object to his "my way or the highway" approach - it's his project, he can surely run it however he likes, and if he wants to be an asshole then that's fine.

That doesn't mean that the rest of the world needs to tolerate him, though, and I imagine it won't be long before he's sitting alone in WordPress HQ while everyone else is running a shiny new fork. Meanwhile, his ongoing hissy fits are a source of ongoing amusement, particularly for people like me who gave up on wordpress years ago.

Blue Origin gives up on New Glenn lift-off, 2 hours into launch window

"Rocket fanciers who were up at 0600 UTC for the opening of the launch window saw repeated resets of the countdown clock"

That would be me. Disappointing, but not unexpected with a maiden flight. No doubt they'll get there in the fullness of time.

SpaceX will try satellite deployment on next Starship test

"Or perhaps less of the Moon bit. Company boss Elon Musk described the Moon as "a distraction" in a post on X last week. His Christmas gift for NASA was to call the architecture of the agency's Artemis program "extremely inefficient" and a "jobs-maximizing program, not a results-maximizing program.""

Well, I can't say he's wrong about that. But not, perhaps, the most sensible thing for him to say of spaceX's biggest customer. Unless, of course, he knows something we don't.

Fining Big Tech isn't working. Make them give away illegally trained LLMs as public domain

As "solutions" go, this is utter cobblers. The whole point of intellectual property laws is to protect the stuff by restricting access to those you can screw money out of One does not achieve that by giving it away.

Of course, if the model isn't actually a breach of those laws, then forcing into the public domain wouldn't harm the IP owners, but then they'd have nothing to complain about in the first place, so it seems a bit pointless other than to give the OP something to tub thump about.

I'm wondering why they're not also going after Google search - that scrapes the web, caches it and squirts bits of it back verbatim with every single query. And they blatantly make money from doing so. But Google bashing isn't trendy, is it? Won't give you a sexy axe to grind and a nice spot on the crusading lecture circuit.

Starlink direct-to-cell is coming to Ukraine

Pootin isn't going to be too pleased, I assume. Still, with 1000s of those pesky satellites buzzing around, he'll have a job to shoot them down.