Posts by that one in the corner

Judge slaps down law firm using ChatGPT to justify six-figure trial fee

Re: Second time in two days ...

> That is, generally, a good thing.

But not a thing that is related to the point doublelayer was making.

Reacting to the context of the conversation would indeed be a useful feature, if it were based on a solid foundation.

Although, what is the conceptual difference between "reacting to the context of the conversation" and simply re-starting from scratch and feeding in the whole conversation again as the prompt?

In other words, "reacting to the context" is really an amazing feature only insofar as they are spending the resources just to keep your session alive whilst waiting patiently for you to type something, rather than dropping it and letting another process use the expensive hardware.

Re: Bullshit

> a fairly new technology

New? In comparison to other computer tech?

Neural Nets predate WYSIWYG wordprocessors. By decades.

The only "new" things here is the scope of the hype, the way that has reached pretty much everyone who uses computing devices on a daily, nay, minute-by-minute basis.

Actually, even that is barely true. As noted, NNs are old and people have been building them and offering them to compulsive computer users pretty much the whole time. Without much success, beyond making money before the bubble bursts (to the point of just-inside-the-letter-of-the-law fraud[1].

There have been - still are - places where NNs have done good - have saved lives[2] - so there is certainly no reason for thinking that there is some irrational anti-Net bias: but those genuine successes are not the subject of the massive hype machine.

The change now is that so many, many more of us are now in the minute-by-minute computer users category and the economy of scale means that spending one dollar per user has allowed a (very small number) of players to build some very big 'nets.

But being big doesn't mean the maths and logic behind these things has in any way suddenly changed[3]. All the flaws are still there[4].

[1] e.g. 1980s/90s, as automated trading took hold: take a stock exchange feed, train up as many 'nets as you can get PCs to run them - don't worry, the data rates are low enough and you don't need to waste 'net nodes on pseudo-parsing natural parsing language - then switch them to output mode before trading opens the next day: ta-da, predictions for the market that day. Repeat for, say, a month. Most models' predictions made losses - but this handful won Big Time! With a totally straight face, sell copies of that handful of 'nets (with witnessed guarantees that they did make those predictions!) to anyone with a big old pot of cash.

[2] 'nets used in medical image screening - *but* we just happen(!) to give very high value to the true positives, enough that we are willing to forgive the false positives - and quietly shrug our shoulders over the false negatives, because we are still catching more than we did previously.

[3] there are good reasons why AI research spent time on more than just NNs - and it wasn't just because the hardware wasn't available: Moore coined his law a while ago (and the current LLMs are appearing in a time where we are concerned that said law is losing steam)

[4] big bug bear: inability to explain their reasoning *and* to have that path tweaked to improve the results; instead, next run (irrespective of whether you change the prompt) you get a totally new output and need to go over inch of it again to look for flaws.

> it makes up nonsense, but it suggested a line of attack he hadn't thought of, which worked.

These anecdotes, where someone - anyone - has used ChatGPT and "it didn't solve the problem, but I came way with a new idea" make me wonder if they could get the same result by using "Rubber Duck Debugging" and a random choice of new approach to the problem by, say, drawing a card from Brian Eno's Oblique Strategies deck[1]. Cheaper, far fewer resources needed.

Now, if you have a citation, including the transcript, so we can see that it was, indeed, ChatGPT that came up with the idea out of the blue, without either being led by the nose[2] or it being one out of dozens of otherwise useless ideas[3] then we can reevaluate the level of incredulity.

[1] hmm, to make it more Modern and Computery, we could drop a copy of those cards into ELIZA and save on buying rubber ducks; I mean, have you seen how much they charge nowadays for a decent model, like a Shakesduck or an Isambard Kingduck Brunel?

[2] I've had some useful info out of ChatGPT, because, unlike Google, after it has come up with the latest, most repeated/hyped, suggestion, you can tell GPT "no" and it'll dig out something a bit more obscure; did that a few times and it finally spat out a reference to some 1970s tech that turned out to be what I wanted to find.

[3] carry on saying "no, that isn't what I want" and it'll happily apologise and then spit out yet another semi-random selection, throwing stuff on the wall and seeing what will stick. Then a simple bit of confirmation bias[4] will get you to remember - or just retell - the sticky bit and forget to mention the mess piled up on the floor or the bits that were so off-target they flew out the window.

[4] even the smartest people are subject to this; I'd even wager the smart ones are more used to having lots and lots of thoughts about a problem that they are really used to just forgetting about all the ones that never lead anywhere that they may not even notice how many such fly past.

"was misbegotten at the jump"

At the "jump"? Misbegotten from the start, surely, at the very point someone on the legal team even had an inkling to use[1] ChatGPT for this purpose, let alone the point where, having prepared everything, they finally made the jump and submitted the prompt.

[1] would say "misuse" here, but that would imply that there is a sensible way for these people to "use" ChatGPT in the normal day to day[2], and it is difficult to come up with a scenario where anyone in the legal profession should ever hand over responsibility for writing anything to ChatGPT, let alone something that is intended to be examined by the one person who has such power over the case!

[2] but what about simple, repetitive, things, like writing flyers to leave under windscreen wipers, I hear you cry? Well, if the interns aren't doing the simple assignments properly first, they'll not be getting the practice needed to get the meaty stuff right. Though if they are in a firm that advertises by interfering with parked cars...

China breakthrough promises optical discs that store hundreds of terabytes

Re: Read/write speed may be the barrier to adoption here

With 100 layers, add as many heads as you are willing to pay for. The rotational latency will remain the same, but you wanted throughput.

AI comes for jobs at studio of American filmmaker Tyler Perry

> Hasn't he heard of matte painting...

You can't expect a man of Perry's stature to know anything about the history[1] of the craft, he is a Going Forwards only kind of guy.

[1] "history" of moving film? Good grief, it is so young that it barely has history! Compared to theatre, say... Next, you'll be telling me people involved in something newer, like computers, don't bother with their "history" either! Preposterous!

Europe's data protection laws cut data storage by making information-wrangling pricier

> The cost savings appear to assume that there was no benefit to holding the information. Doesn't sound likely.

If you're not concerned with any privacy implications, it is (initially, at least[1]) dirt cheap just to grab every bit of PII you can and hang onto it, rather than spending time & effort trying to decide if there really is any benefit to you from hanging onto that data: it looks juicy and interesting, we'll think of a use for it.

A lot of stuff, of all sorts, gets hung onto without any benefit, or the apparent benefit evaporates. Haven't you ever been a part of the occasional "emptying of all the cupboards and dusty storage bins" in the office/factory?

[1] as time passes and the company/clients/customers increase the unnecessary information can take up enough space to become notable - by which time it continues out of habit, assuming *someone* knows why it is being kept around.

China could be doing better at censorship, think tank finds

Was starting to feel sympathetic towards the censors

The article painted an all too familiar picture: a large organisation, nobody seems to be actually taking overall control (but are basking in the claims that they are in total control), resources being spent over here to duplicate effort whilst over there is lacking the necessary to get the job done.

Then was finally reminded at the end, nope, just for once, this is what we *want* to have happen, we don't want an efficient censorship.

Still have a bit of fellow feeling for the grunts at the bottom, just trying to get the day's work done, despite it all.

Orgs are having a major identity crisis while crims reap the rewards

Social engineering, the second oldest profession

Making phone calls, convincing people to bypass the mechanical security...

Have we become so focussed on the clever coding and machine based security that we've been forgetting to train everyone to look out for the old tricks? Human Factors Engineering, if you will.

(Ok, some of it is just good old fashioned laxity - not using your TFA - which is just the back-end Human Factors kicking in)

Neuralink patient masters mind-mouse maneuvers – if Musk is to be believed

Long term care after the tech is declared obsolete

> However, there is no information on what might constitute an adverse reaction nor how the implant would be removed if required.

We are rightly concerned about how Neuralink are going to behave if things go wrong, but we should also be asking about their plans if things go right.

How long are these implants intended to work? What will happen when they reach their natural end of life (physical breakdown of parts)? What about when they are declared obsolete and are unnaturally declared EOL by Neuralink - having moved onto the Next Big Thing with a cohort of Bright Young Minds who have no idea how the existing system works?

Are patients going to receive an email saying, sorry, but in the next hour your ability to communicate with the outside world is going to be switched off? You did keep up practising with the clunky old pick-a-word board, didn't you?

The recipients of Second Sight's visual implants have been through this - are there signs that Neuralink (or any other commercial implant company) are taking their long-term responsibilities seriously?

At a bare minimum, placing *full* details of the device and support systems into escrow, in hopes that someone is able to at least try and help the victims.

masters mind-mouse maneuvers

I was relieved when the first sentence pointed out that this is referring to a *computer* mouse.

I had visions of Neuralink just connecting the human patient's device to the lab wifi and wondering why the caged mice were suddenly so attentive - and why the patient was asking about cheese.

Superapp Gojek fine-tunes each new error message for a week. What? Why?

"motion designer" is in some ways just a new way to say "animator."

> The difference is that the format and style has changed over the years.

It has changed enough "over the years" to warrant a new name, something that "animator" just doesn't conjure up. Hmm, interesting, tell use more.

> In the past, animations were heavy on being verbal explainers. "They used to have a character and say things like 'this is Bob …', but they don't do that anymore.

Why does this sound suspiciously like a description of low-effort early-morning children's TV cartoons? Do you think that is their entire reference pool for the work of animators?

> Higher-end studios prefer something more abstract," he explained.

What? What! They prefer something "more abstract" than anything those olden times (/sarcasm) animators turned out?

Something more abstract than Prelude Taking A Line For A Walk. One of my personal favourites, the marvellous Erica Russels' Feet Of Song and Triangle (you may remember her style from a number of UK TV ads and some music videos back in the 80's and 90's). Or go and dig out some of the Drawn-on-film animation such as the nobody-ever-saw-it (/sarcasm) A Colour Box ad for the GPO that was shown in cinemas across the UK (i.e. these abstracts were never intended, nor treated, as some niche product, hidden away from the masses - which makes it all the stranger that someone who "knows enough about animation to realise what they are doing now is so very different it warrants a new name" seems so ignorant).

> A wordless animated image may also be a clever way for a business that operates in five different countries

"Wordless" - well, although all the above integrate the sounds with the visuals, they are wordless. Okay, the dance pieces work do rely on the soundtrack, so how about something from Jan Švankmajer: Passionate Discourse - you could even use part of that as a progress bar for, say, a diff/merge operation!

All of the really great animations work without narration or text (ok, the aforementioned A Colour Box has some text in it, so I probably should have left it out - except it is such a nice example of dissemination of the abstract; ok, ok, I like it and want to share it). If you switch off the sound and just watch the narration-heavy The Man Who Planted Trees you are still shown a story with a depth, from the abstractions, that a straight filmed version would struggle to capture (sorry, sorry, getting a bit into film reviewer mode there).

> "It goes back to ten to twenty years ago when we surfed online and there was nothing to tell you if a website was loading. Then they introduced a preloader – a bar that progressively was filled in, so visually the user doesn't think the browser is hung," mused the creative director.

As Dan 55 pointed out, this "creative" director has missed *another* piece of history, the 30 year old browser progress indicator (and still that ignores the pre-browser progress bars, let alone the even older "print out a dot every now and again, to let the User know we haven't crashed" precursor to the Mosaic planet icon).

So not only do we *not* need 'a new way to say "animator"', attempting to do so shows a total lack of awareness of the subject (gosh, some of those references are almost 80 years old, they can't possibly be of relevance to Today's Exciting New Apps) but is actually insulting to those people who pioneered the methods.

PS

For more, start with the 4Mations YouTube list.

PPS

Once more, I am merely a programmer, not a Creative, let alone a Creative Director, so what do I know; we operate in totally different areas of work and I can not possibly expect to hold them up to the same requirements about depth of research or general knowledge that are needed for the boarish work I've done to earn a buck.

Chunks of deorbiting ESA satellite are expected to reach the ground

Live and Direct

The Zik Zak corporation turned this sort of thing into an annual festival, Sky Clearance Day: bring your best metal umbrella and join the fun!

Max Headroom, prr-prr-prrredicting the future - well, 20 minutes into it, at least.

Re: If you'd asked me ten minutes ago...

Don't know about the Spinal Tap guy, but Maggie O'Connell's boyfriend was killed by a falling satellite in Northern Exposure - that was a particularly hardy piece of kit, as it still had its antennae intact at the funeral (they just don't make 'em like that anymore).

Forgetting the history of Unix is coding us into a corner

Re: Man pages

You had enough space to keep all the man pages (unformatted) online all the time? Luxury!

If only Ikea had expanded internationally earlier, they would have made a killing with a Billy perfectly sized for the ring-binder man pages.

Attempts to demolish guardrails in AI image generators blamed for lewd Taylor Swift deepfakes

Re: why is nobody...

And the things they've done to ducks. Oh dear heaven, the ducks!

Developer's default setting created turbulence in the flight simulator

Re: sort of on topic...

Was it one of those highly educated sorts who somehow manages to get themselves into endless scrapes on an almost weekly basis?

OpenAI reassures: GPT-4 gives 'a mild uplift' to creators of biochemical weapons

One step closer to taking over the world

> OpenAI compared results produced by the two groups, paying close attention to how accurate, complete, and innovative the responses were.

Accurate? Complete?

OpenAI already has people on-hand who are expert enough to judge the accuracy of instructions for creating bioweapons?

Or did they just work through all of the answers and see which ones gave working results on their tests groups?

Either way, are OpenAI taking the Ernst Blofeld approach to Market Domination?

Microsoft seeks Rust developers to rewrite core C# code

Re: Rockstar

> Maybe the move is to attract young, brave, naive, rockstar developers

Or to attract grey beards who have been pissing themselves with laughter over C-octothorpe since it appeared, by replacing it with something that demonstrably helps with memory management.

The literal Rolls-Royce of EVs is recalled over fire risk

Re: Required edits:

> British marque's' -> 'German marque's

Cough.

British brand's' -> 'German brands's

As neither of those are France.

The marque that sparks

This isn't just any EV...

Zen Internet warns customers of an impending IP address change

> most of our customers primarily use just one IP address

And how many of their customers only *have* just one IP address?

Like those who joined after Zen were well established and had stopped giving out multiples as an incentive...

Re: Hard Work

> sorting out their situation with Zen or moving to a new provide

And we go around and around again - I changed to Zen in the first place on the advice of El Reg comments

We only have the one IP and haven't (yet) received this notice, so fingers crossed, but with the tales of customer service degrading...

Then again, with OpenReach seeming to have taken the government's deadlines for death of POTS and supply of fibre as a way to push back providing fibre on our exchange, there is time to take a breath before the next forced change to our service and contract "upgrade".

40 years since Elite became the most fun you could have with 22 kilobytes

Re: Better than Dangerous

I backed out and took the refund when they dropped standalone single-player mode and insisted it all had to be massive multiplayer online or nothing. I know, very much not down with da yoof and all that, but I'm quite happy to play video games only sporadically and take my own sweet time to progress (one day I may even finish Riven, no hurry).

I was quite annoyed about that change (as an original BBC Elite player I was looking forwards to the new sparkly bits and pretty pictures) but at least they did refund us.

Elon Musk's brain-computer interface outfit Neuralink tests its tech on a human

Re: This could be a really big deal

> I could also see fighter pilots and the pilots of Musk's Mars ships connected directly into their crafts[1], no joy sticks, screens or any of that, direct input and control.

No screens?

No *screens*?!

You want this thing to pump data back *into* the pilot's mind? Forget Firefox, worry about Firewall Down!

Yikes[3]

[1] their craft! The plural of "aircraft" is : "aircraft"! Unless you were thinking they would learn macrame and the carving of scrimshaw[2] whilst the craft's FSD flew them (in)to Mars?

[2] perhaps a scene of a privateer with four or five cannon along the port side.

[3] at this point, trying to list all the synonyms for disbelief suitable as a response to this, my cortically implanted thesaurus overheated and shut down and just emitted a single "yikes" before emailing to schedule a service.

Linus Torvalds flames Google kernel contributor over filesystem suggestion

Re: I am Ignorant.

> Off the top of my head, find and tar need unique inodes

True. And if you read a few of the messages, Linus accepts that.

What he isn't accepting is that there is any good reason to support someone running tar over the particular FS in question, eventfs (and/or tracefs, the distinction is a bit confusing). Nor does he worry about people having tar fail when trying to archive procfs! If you try to tar eventfs, the argument appears to go, even if it "worked", what good is the result?

The additional argument (over inodes that will about allow for processes in tight loops doing mkdir/rmdir all the live long day) is - more than I want to worry about! Came back and read some more Register articles instead.

Fairberry project brings a hardware keyboard to the Fairphone

Sharp Zaurus

Purely for the nostalgia, I just picked up my Zaurii, pulled down the slider on each and revelled in the feel of the keyboards - the sleek silver of the 5500 and the bold black 6000: so much easier than a touch keyboard and far less likely to just slip from the grasp. A decent single-handed thumb typing experience (and being able to use a stylus, just like the Palm Pilots, made for simple precise pointing as well, but that is another discussion, about the cost of a Wacom-enabled 'phone...)

I hear tales that a modern 'phone can be used with the thumb of the hand holding it, but I have yet to manage that without speaking out loud all the gt£%@4ghy%)'$°] that inevitably results.

The pen is mightier than the keyboard for turbocharging your noggin

Re: Why aren't you taking notes?

> I have to exercise and practice many times to get things to stick.

Aside from the simple "horses for courses" and "diff'rent strokes" from person to person, you are also talking about needing to put different amounts of effort into different parts of the process.

In other words, "taking notes" and "exercise and practice" are not in conflict with each other but act in different ways (and with different strengths between individuals).

So my experience differs from yours: if I am in a lecture and don't take a continuous stream of notes then the words won't even enter short term memory, but will tend to just wash past me; the act of note taking engages focus (for me). Otherwise anything else going on may capture focus (the people nattering in the audience, what that person is doing on their phone, the architecture of the lecture room).

Once past the hurdle of focus, the info bounces around short term memory so the writing can be notes (in full sentence form) rather than simple dictation.

Reviewing the notes provides the chance to see if there are any questions left at the end (and maybe the chance to ask them).

But if I then want to use the new information later on, I still have to practice - looking over the notes again in the following weeks, applying what they say to exercises, comparing it to other material in related areas. Drilling it into long term memory and tying it in to everything else (if it can't be tied into other learnt material then it isn't going to stay for the long haul - don't ask me to memorise a list of kings of England).

We put salt in our tea so you don't have to

Re: Ramble, blither

> Most butter in Tibet is rancid

According to my in-depth studies of an old Journal Of Record[1], the supply of Tibetan rancid yak butter has been endangered by the import of modern detergents, as the yaks are no longer rancid.

[1] The Beano

What is Model Collapse and how to avoid it

What steps should the machine learning community take?

A: stop blindly scraping everything off the Web.

Even without seeing your own output again, scraping everything was never going to provide a totally sane dataset in the first place.

But it was cheap and easy, because the only way they know to build these models is to keep on shovelling into the training maw.

But labelling, manually attaching provenance and "truthiness" values is expensive (and gets away from the Tech Boys Building Bigger Machines), and becomes unfair for smaller companies.

Shame they don't have a better way of building their models.

Akira ransomware gang says it stole passport scans from Lush in 110 GB data heist

Passport scans

> are routinely collected to verify identities during the course of the hiring process

Collected? Not just "checked and that check signed off by HR"?

Oh no, of course not, that might mean that HR actually has to take responsibility for doing their job and putting their name to it. Far less damaging[1] to demand and keep a scan.

[1] to that HR hack, that is.

At last: The BBC Micro you always wanted, in Mastodon form

Re: BASIC

> Other languages I saw were Comal and Lisp, and I'm sure there were more.

Had a few happy hours playing with microProlog on the Beeb.

NASA engineers scratch heads as Voyager 1 starts spouting cosmic gibberish

Re: A repeating pattern of ones and zeroes

> so the fabric of space could be gingham with a houndstooth pattern

That happens often enough in Science: two competing hypotheses battle it out for years, then the observations reveal that Reality is a combination of both.

A repeating pattern of ones and zeroes

Having reached The Edge, it is trying to send back a picture of what it has found there: the Fine Structure is revealed to be almost a black and white chequerboard. Theoreticians are waiting with bated breath for the full release of the data, which will be pored over in detail, looking for the slight changes in the pattern 0,1,0,1 which will finally answer that hotly debated question:

Is The Fabric of The Universe gingham or houndstooth?

You don't get what you don't pay for, but nobody is paid enough to be abused

Re: Question

> I've been skim-reading it this entire time as "Regonomize"

That is the best approach to reading this web stuff nowadays: let your brain feed you the *sensible* words

I mean, come on: yes, you can slam two words together to get "Regomize" but to defend *that* as being the better - or even "correct" - result, you may as well go the whole hog and use the word "Rize", or even "Re"!

Although, cough, "Regonymize" would be the simpler choice, no need to make weird claims about creating portmanteau words, it just says exactly what El Reg is doing - picking a name to fit the El Reg style, the same way that "anonymize" says precisely what *it* describes.

PS

Apologies for all the zeds but one battle at a time.

Microsoft Forms feature request still not sorted after SEVEN years

Re: What kinds of questions will need time input and what follow-up actions require the time?

> Microsoft senior manager of a product intended for business use could not possibly imagine a scenario where a business activity needed to have a time associated with it

Time? Be glad he didn't ask if anyone *really* still needs a Date field.

Because whenever anyone asks a senior manager when something is needed, they always reply "yesterday"!

What's the golden age of online services? Well, now doesn't suck

> Hand up, I bet quite a few here were on Demon's "tenner a month".

Sir, me, sir. On the list with my prepayment, so he could get it set up in the first place.

Not forgetting hacking away at Phil Karn's KA9Q software to streamline it a bit for DIS, then it gained a termulator, TextWin[1], that let you do more than one thing at a time. Under DOS!

[1] glad you reminded me of DIS now, as the bottom of that URL (the first hit for TextWin) lists one John Washington as the author of TextWin: he appears to have taken on the maintenance of TextWin (glad someone did), but Anthony McCarthy (sadly no longer with us) was the original author, as noted here (buried in a long chunk of text). I helped test & debug TextWin with Tony, so I've made myself a note to go and have a proper look at what became of it.

Re: Not an angry place .

> I'm so fond of the old 56k modem handshaking sounds I've had it as my ringtone in the past.

I have it on a tee-shirt.

> people would pick a service on the basis of their file library and how much it cost

Ah yes

> per hour to download them from the online service to your PC

*or* how much they would charge to pop a floppy into the post with Volume 11 : Best of LaTeX on it! Far cheaper than downloading - and it felt a lot faster as well (no babysitting the connection)!

> BIX, anyone?

Nobody ever got raked over the coals for running up an obscene (really, really obscene) 'phone bill from sitting on BIX well into after-office hours. Nope, didn't happen.

> Commercial Internet eXchange (CIX).

That was a Great Thing. As well as reducing the 'phone bills compared to BIX (!) and having forums for properly local people to share news in (Bristol, Avon not Tennessee!), they released the offline reader Ameol, which made it even more affordable - and I still use Ameol as my primary email client[1]

[1] Ameol - it does everything that I want it to do, avoids things that I want to avoid - and the only time that it seemed to have gone bad and refused to show my emails (it is a Win32 exe and can not index into archive files bigger than 2GB - or was it 4GB?) one simple use of "Purge anything you've explicitly marked for deletion" and all was well. They just don't make them like that any more.

Digital memories are disappearing and not even AI or Google can help

Re: You have HUNDREDS of browser tabs???

> I don’t understand why you would have more than one or two tabs open at a time

Right now, I'm working my way through talking to some sensors across I2C using Python - nothing World Changing, but I've not done it before. I have tabs open for the MCU specs, the sensor specs, the pin assignments for the dev board, the pin assignments that the software expects, the tutorial that I'm working through, the example code that tutorial is working towards, some more example code that relates more closely to the dev board I *actually* have, multiple pages in my personal Wiki so that I can quickly add a sentence - or a URL, with all the context that a "bookmark" totally fails to provide - about each item as and when I realise that would be useful. Plus a diary/logbook Wiki page into which I note down all the commands and actions I *actually* do whilst following the tutorial, because, hey, I'm fallible, I'll miss out a step by accident and need to backtrack. Oh, right now, I also have open The Register, as I'm taking a break, and sometimes a concert video in *another* tab, as I like a bit of music while I work.

One or two tabs? Heck, I have these spread over multiple *windows*, both for convenient grouping (boring old specs sheets stay together) as well as to allow the logbook page to stay open on the portrait monitor.

I guess I *could* keep closing and re-opening each tab, but that doesn't exactly make for a smooth workflow.

None of this is meant to be boasting - as I said, I'm not doing anything groundbreaking here, just playing with a sensor (and, soon, blinky lights! Yay!) and, finally, finally, getting around to learning a tiny bit of Python, like the Cool Kids do. But hopefully it illustrates that having more than two tabs open *can* be productive. This is all the sort of thing that I do *now* to keep on top of stuff - and I *really* wish that I'd figured it out (and had the physical means to do it) e *lot* earlier, when I was still working, especially in the early days. So much stuff would not have been forgotten - and would not have had to be re-implemented purely because it had been forgotten.

> Why not just bookmark stuff?

Because the browser's bookmarks list is utterly dreadful for organising and providing context for the URLs; no, adding folders of bookmarks isn't The Answer: which folder does that really belong in? It is great for keeping track of a dozen or so items, like the *correct* URL to keep up with The Register, the easiest way to read XKCD or the homepage for remote-controlling the DVR[0], but the URL for the tutorial I mentioned above is, for my purposes, important in at least three distinct contexts (two for the hardware it uses - and one just for "this is what I did in 2023"!); expressing that in a bookmark list just won't fly[1].

And by the time you've bookmarked a hundred items or more you are starting to spend time just searching for the URL in that list. Assuming you even remember that there is anything worth searching for.

You could start to edit your bookmarks, removing things that, say, you haven't looked at in six months. I won't do that, because, if this sensor and lights thing works then the hardware will last for years - but I'll want all the records about how it works, why it *doesn't* do "something, when asked to "make it do something to be different last time".

> Who cares if a 30 year old file isn’t readable any more?

How about putting together a memory board (and in case you want to dismiss those, consider funerals and dementia) - and you are handed a file from a distant, and thoroughly non-techie, relative of the memorialised, who happened to remember he had this sitting in an old directory buried on his hard drive: it is all the notes from when we got together 30 years ago and first had the ideas which she surprised us all with when she turned them into that book, the one that made her World Famous in Richdale. That file obviously means bugger all to you, Dexter, but now they know it still exists, it would mean absolutely everything to that group of people.

[0] grr - El Reg won't allow an href to 192.168.254.40 for that DVR homepage; it is obviously used to commentards trying to be "clever"!

[1] which is why I use a little personal Wiki: I can add a note about *why* I've saved this URL. I can add it into lots of different places, I can reorganise at my leisure. Other solutions for organising URLs are available, even websites (which are just single-use Wikis, aren't they? With the advantage that you are sharing all that information about your interests with the website operator).

Re: ...And Then There's The Problem Of Old Application Software....

> TIFF (yes even TIFF) ... TIFF is a container for a whole lot of options that you can shove inside and still give it a .tif extension.

Aaaaaarrggggh TIFF!

I have long, long been of the opinion that practically nobody who suggests anyone use TIFF can even say what it name expands out to, let alone what data formats it can contain! Tagged Image File Format - and even if Aldus (oops, Adobe) still play gate-keeper to the list of "known" tags that won't stop some clever-dick from reading and acting upon the Wikipedia article:

> However, if there is little or no chance that TIFF files will escape a private environment, organizations and developers are encouraged to consider using TIFF tags in the "reusable" 65,000–65,535 range. There is no need to contact Adobe when using numbers in this range.

And we all know what happens when you create data that isn't intended to "escape a private environment", especially if a salesman spots you using it..

Even in the last year I recall seeing someone urge the use of TIFF "because it is uncompressed" - sob, bang head on desk.

Sorry, rant over[1] - and sorry for that list scraping over your scars.

[1] Go on, tell us about XPS, we know you want to :-)

Re: ...And Then There's The Problem Of Old Application Software....

> I do not particularly want to pay writers of modern software to implement compatibility with file formats that never show up

Neither do I, which is why I said:

>> providing a way to dump to structured text (or just publish the damn format!)

That is, when you are about to drop a file format, complete the job and dump the information you already have onto the public (especially easy since the rise of public repos - SourceForge to GitHub).

Of course, you can argue that they never had that information to release in the first place - and I wouldn't doubt that many are in that position[1] - but then that gives you another data point to access whether you really want to ever again rely on someone with that level of competence.

> Yes, this annoys me, but not as much as it seems to annoy you... People posting here know perfectly well which formats are open or not and they knew that when they created those files.

As I've said before, "People posting here" are the minority, the Cognescenti, the Privileged Few - and lording your Privilege at the expense of all the other poor sods, the Ghastly Hoi Polloi who have just been dragged into this situation through no fault, and certainly no desire, of their own is the position that *actually* annoys me.

[1] I have spent too many working days implementing dumps because other people in the same group haven't bothered to and, guess what, during code changes errors have crept into code used to write the data! Surprise! So we have to hunt down what those errors are (dump the data to text - and grep like your life depends upon it!).

Re: ...And Then There's The Problem Of Old Application Software....

> I am surprised to see "XML" in that list...useless without the metadata...

Indeed.

Which illustrates that this area is still a work-in-progress.

If enough of us, politely, pass on these observations to the people drawing up those lists, hopefully things can get better (although my contact in the DPC does a very good eyeroll when I get to ranty about the whole subject of metadata/DTDs/schema).

Re: ...And Then There's The Problem Of Old Application Software....

> I suggest that this problem can only get worse over time.

Sadly, it will. And the impact that has is never considered by those responsible.

Archivists are concerned about this and try to publicise the problems, for example Recommended File Formats for Digital Preservation, but not only us the fight against inertia and "that is your problem, not mine, why are we paying you archivists?" but look at the list and notice the lack of "whatever is Word's current format" under word processing. Now, we all know of the news stories where one body or announces that it will no longer use proprietary formats when Microsoft et al swoop in and spread the FUD (heck, look back over suggestions in Register comments that we'd be better off using Open/LibreOffice than MS Office and spot all the claims that that just isn't feasible, being polite about the arguments against).

And that is just (!) when talking about Very Important Documents, ignoring all the data lying around which Society (and certainly Big Important People) considers trivial, but actually has a big impact on people's lives and is therefore actually worth preserving: digital entertainment, such as games. The death of Flash and the creation of the Flashpoint Archive is taken seriously by the Digital Preservation Coalition. Meanwhile, all sorts of bits of information are going missing because the people who *could* make it preservable, the software authors[1], consider it their right to be able to (effectively, and sometimes literally) chuck it away.

"What", you say, "literally chuck our information away? Nobody does that!". Have you checked everything that is sitting on Someone Else's Computer?

Records of interactions that you sometimes want a reminder of? Of course *you* religiously download your own copies of everything you might want to refer back to - but then you realise the email invoice just has the part number on it and you have all been relying on the online system's copy to have hyperlinked that part number to the full description page, until their sales system was Improved and all those links died: you still have the data but have just lost the means to interpret it, without which it is just so much noise.

And in all of this, Register readers are the Privileged Ones: we know that They do these things to us and (hopefully) take the time and trouble to take appropriate measures.

Meanwhile, the poor unsuspecting in The Sea Of Users out there are gambolling in the shallows, unaware that they merely on the top of a sandbank that can suddenly erode away beneath them.

Re: ...And Then There's The Problem Of Old Application Software....

> Yes, that will happen

Gawd, I hate that attitude: the creators can't be bothered to keep their own software compatible and your response is to shrug your shoulders and lay the onus squarely on the user: "well, if you *cared* about it, you would have bought every update and made sure you spent countless hours loading and re-saving every file".

And, of course, that should apply to *every* *single* *user*, no matter how techie they are - or aren't.

The creators who simply can't be fagged to support their own file formats, not even to the point of, say, providing a way to dump to structured text (or just publish the damn format!) so that everyone has a fighting chance of recovering old files - those are really annoying.

But that kind of "roll over and play dead for the pretty creators" attitude is simply loathesome. And that is when the attitude is pointed at other commentards! We Readers of The Register are the lucky ones, we are the ones that *know* creators will simply abandon us. To expect - nay, demand, else they clearly don't care - that understanding from all the Users out there is simply beyond the pale.

Re: Preserve the meaning of our personal past

> People who hoard physical items, filling their homes to their own exclusion, are rightly considered to need help

Between ascetic minimalism and the point where hoarders literally exclude themselves from there own homes there lies a very large range of options. The same behaviour can lead to dramatically different outcomes purely due to circumstance: the avid collector of foibles is either an interesting old duffer, whose glass cases of oddments are preserved for future visitors to his country house or a leathery corpse buried under piles in the bedroom of his two up, two down mid-terrace.

With modern storage, we can all choose which route to take: and we csn apply as much organisation as is useful & comfortable to ourselves, right now: if you would *like* to be able to find that photo of the funny kitten, you can decide how much effort it is worth it to yourself, today[1].

But for the interests of the future:

> but most of the rest will evaporate with my own demise and if it were to vanish sooner it would not be a tragedy.

Find yourself an archaeologist and ask them to talk to you "middens". Take a packed lunch, you will be there a while.

The illregarded and the discarded are what paints a picture of everyday life: our future researchers won't be getting that from carefully preserved presentations from the movies or TV (if every pub exploded with the regularity of the Old Vic we'd need flack jackets to safely visit the village corner shop!). In terms of bulk personal digital storage, "discarded" is not so much the contents of the Windows Recycle Bin that you never bother to empty, it is more the files you put down, can not for the life of you remember where they were, but there is plenty of space left, just forget about them.

Even if we haven't got the software tools to automagically organise and make sense of all this stuff today, I think we can safely act as though it will become available, one day.

For now, we just need enough to keep ourselves supplied with virtual Victorian glass display cabinets - and have a Serious Think about how & where to keep all our old NAS devices so once we've popped our clogs they aren't just wiped.

Maybe the future would *want* us to keep putting PCs and NAS drives into (designated) landfill, just so that they know which techno-middens to dig in once they have perfected the datamatic trowel and measuring stick.

[1] FWIW these days I seem to spend up to day a week - being retired - just adding cross-references and filling in blanks of "things that I won't forget, no need to write down" in my personal Wiki, which has been keeping track of things for me since the late 90s; all just to make it work for me when I do want to recall something. Oh, how I wish I'd dome more of that cleaning up as I went along.

Apple and some Linux distros are open to Bluetooth attack

>> most Android users are on their own as the OEMs get bored of updating their phones so quickly...

> I understand your angst. But I am very trailing-edge, yet I looked around and my only 'fone with <=10 is...

Huh? How is that a response to the comment you quoted from? So you received a message about one particular bit of software which will no longer support Android 10, what does that have to do with OEMs not bothering to pass on updates?

Were you perhaps reading the statement from the article that "Fixes for these issues that affect Android 11 through 14 are available to impacted OEMs" and believing that meant all the OEMs actually *will* update *every* device that has versions 11 upwards on them?

'Cos, nope, Google having a fix available is nice and all, but it has bleep all bearing on the OEMs actually pulling their fingers out, applying the fixes and delivering them to anyone other than people with a 'phone that was new to market less than, ooooh, five months ago, that is about how long we can expect their interest in us to last!

PS this tablet is stuck on Android 6.0.1 so guess I'd better never turn the BT on again!

Boffins devise 'universal backdoor' for image models to cause AI hallucinations

I'd be chuffed to bits if someone honestly referred to me as a boffin!

BONGO: So, the lab boys have come up with a drive that can break the speed of reality.

ACE: Those boffins have hammered together a crate that can cross dimensions? When do I launch?

Can't see the wood for the trees

> There are various possible attack scenarios ... posting a number of images online and waiting for them to be scraped by a crawler, which would poison the resulting model given the ingestion of enough sabotaged images.

> "Where these attacks are really scary is when you're getting web scraped datasets that are really, really big, and it becomes increasingly hard to verify the integrity of every single image."

So:

* you are scraping the web for damn near every image you can lay your hands on (there will be some filtering applied - don't bother with the 1x1 tracking images, for example - but broadly speaking, every image).

* some sort of data within an image counts as "poison" and you only need a very small percentage of that to ruin your dataset

* up until now, nobody knew what this "poison" looked like, so nobody could have been on the lookout for it[1]

* random images scraped from random places *will* have random data in them[2]

* Now, what are the chances that some of the images you have all already consumed aren't accidentally, randomly, poisonous and all your datasets are already deadly if eaten in excess?

For that matter,

> hard to verify the integrity of every single image

if you're taking arbitrary images, what on Earth does it even mean to say you are "verifying the integrity of the image"? An image can be anything, absolutely anything at all, that is sort of the point of images, and expressing art in your images!

If, on the other hand[3], your "integrity of the image" *really* just means "a malformed file, the PNG standard says that byte should really be..." then you aren't talking about "poisoning images", you are talking about "your image parser is rubbish, what, did you copy out of the UEFI code?" in which case the scrapers should be subjected to Howls of Derisive Laughter, Bruce.

[1] even assuming that they *can* be on the lookout for it - I haven't read the preprint properly yet: it is possible that this "poisoning" is like some of the old antagonistic image changes, where they just fuzzed the image until the classifier puked, but never actually understood what was going on inside the the blackbox 'Network that upset it so. So those trials didn't provide a description of what was important about the changes and therefore what you could do to look out for similar "poisons". That is, they had a demonstration as Proof of Concept "you can ruin the results" and that was it.

[2] You say that I've poisoned that photograph, I say I just went at it with GIMP for artistic effect.

[3] As I said, will get around to reading the preprint, but right now...

Veteran editors Notepad++ and Geany hit milestone versions

Re: EMACS or death

> IBM CUA didn't adopt them till later

Rather, IBM gave up on CUA and didn't tell anyone.

CUA: Cut is SHIFT-DEL, Copy is CTRL-INS, Paste is SHIFT-INS

Windows does [1] honour CUA keystrokes (just tried those three in this here browser edit box) , but if you are using CTRL-X, CTRL-C and/or CTRL-V then you aint usng CUA.

Liam may well be using CUA, but I'd venture that is a dying art (says the person who admitted to using CodeWright and its decidely non-CUA, non-Windows-common-usage keystrokes: Copy is GREY-+ (on the numeric keypad) and always will be!

[1] although not necessarily all programs under Windows!