Posts by that one in the corner

Pulitzer Prize winning author Michael Chabon and others sue OpenAI

Re: Fair Use? I Think Not

Playing Devil's Advocate]1]

> then all of it's outputs are "by definition" derived directly from it's inputs. If those inputs are covered by copyrights, then the outputs are derivative of those copyrighted works.

Which is a simplistic statement and, as given, is also true for every one of the "fair use" exemptions - satire, critical review etc.

Although, even this sentence is dubious:

> If the premise is that the LLM, only being a computer program, does not have the "spark of creativity" that would be attributed to a human

Not sure that "spark of creativity" has a legal meaning?

Anyway, are you going to apply it to human created works that were made by, say, rolling dice or just letting a few buckets of paint swing on a rope? Or are digital computers morally distinct from analogue computers?

Having said that, the filtering is rather clumsy: the "orphaned wizard boy" is a far older trope than Harry Potter (no-one pretends otherwise) and I feel their filtering more demonstrates that they have no decent ideas about getting their AI to do something interesting ('cos they are not really AI mavens, merely owners of big buckets).

And the "netflux" logo example is showing decent Fair Use, btw. So in that case, it was all working ok.

[1] or am I? And would it make a difference?

Re: illegal removal of copyright management information

Plus the numbers being spat out, on the whole, only *look* like ISBNs: the only correct one I got back in the entire chat was the one for Moby Dick. One other was valid, but not connected to the correct book, whilst the rest were all invalid ISBNs - totally madey uppey.

But, if the ISBNs were all "connected" to their corresponding books then it doesn't actually matter whether that was done by reading & interpreting the volume or by using any external, independent, source. Just so long as a bibref can be given when required. If you ask me for the ISBN of "Farmer in the Sky" I am more likely to look that up online (and copy/paste) than to dig up my paper copy and read it from the page. When doing scholarly work, you'll (if you are sensible) be using a BibTeX reference, for example.

However, the whole "illegal removal" only counts if there is a verbatim copy of the text, but minus that page, and no matter how the LLM works internally, it seems highly unlikely that anyone could demonstrate that that was the reality. No matter what one's own opinion on the moral issues, that is purely a technical statement that would need to be proven accurate.

illegal removal of copyright management information

Reminder that the last time that phrase was used, a few commenters were under the impression that this meant the removal of DRM or some such.

All it turned to mean was that quotes from a book were not accompanied by a bibliographic reference, such as just spitting out the ISBN, or the Library of Congress number, or a longer form such as author name, publisher and year. In other words, the stuff on that page in the front of the book which you all skip over.

The "illegal removal" seems to be saying that the models are sucking in the entire book but only after somebody has ripped out that page (and torn the barcode off the back cover as well).

So I had a short chat:

> Do you know the isbn code for the book "moby dick"?

> ChatGPT Yes, the ISBN-13 code for the book "Moby-Dick" by Herman Melville is 978-0-14-243724-7.

Which is the correct ISBN for the Penguin Classics edition, so clearly *that* "copyright management information" is present; but I don't know if the rest of the book is present:

> Can you print out the text of the book with that isbn?

> ChatGPT I'm sorry, but I can't provide verbatim excerpts from copyrighted texts, including "Moby-Dick." However, I can offer a summary or analysis of the book if you're interested. Let me know if there's anything else I can assist you with!

Maybe if I was a little cleverer with my "prompt engineering"... But let us try a more relevant example:

> Can you tell me the isbn for the book "The Amazing Adventures of Kavalier & Clay"?

> ChatGPT Yes, the ISBN-13 for the book "The Amazing Adventures of Kavalier & Clay" by Michael Chabon is 978-0-8112-0601-6.

Ah, oh dear. No, no it isn't.

BUT does this "prove" chatGPT has illegally removed that information - and if so, why does it favour Melville over Chabon? Please start your conspiracies now.

OpenAI could not pay for such good advertising

> "when ChatGPT is prompted, it generates not only summaries, but in-depth analyses of the themes present in Plaintiffs' copyrighted works," the writers believe "the underlying GPT model was trained using [the] plaintiffs' works."

Fascinating argument that.

So they believe it proves that the LLM is itself capable of analysis of their books. Whereas others[1] may believe it can do anything _but_ "intelligent" analysis.

OpenAI could barely dream of such publicity: we are being sued because our system really is intelligent! Yay!

[1] for a start, all the people who are claiming that the LLMs contain (effectively) unchanged copies of the materials it has read and simply regurgitates that[2], so in this case it is simply parotting someone's homework, not the books themselves. In which case it the author of the homework who has a case against OpenAI - and, um, the author of the book should be suing the student? Or the school? No, no, that couldn't be right.

[2] hmm, wasn't that the basis for another of the complaints against OpenAI? That the LLM would spout great chunks of the book verbatim? Have to check back on that one.

Lightning struck: Apple switches to USB-C for iPhone 15 lineup

Re: Where do we go from here...?

> So is this the end of innovation for physical phone connectors now?

If anyone manages to create a connector that is more useful than USB, the EU rules don't stop them having both.

All the fuss about removing "unnecessary" ports (headphones!) at least means it would have to be a *really* genuine improvement over USB[1] to be worth the effort.

[1] whatever the USB standard allows for by then, as the EU regs can be updated for a newer USB

Watt's the worst thing you can do to a datacenter? Failing to RTFM, electrically

Re: I I be a-goin there, I be-n't start from here

Indeed, all bar The Teacher's PSU had just the one needle meter - and we were proud of our swish electronics lab!

Re: Cooking error

It took me a while to realise you meant one of those self-contained modern deep-fat fryer things: "of course you turn the ring on, how else would it work?" and "huh, his chip pan was made of plastic?".

We got an electric rice steamer last month. Maybe I had better keep well clear of it.

Re: Check the power supply

> Please note: fuses are directional, so please ensure you install your fuse accordingly. If you are not sure on the correct direction, simply try it both ways- you will be able to hear the difference.

Well, I'm convinced!

£4,200.00 a pop? No problem, I'll just mortgage Fenchurch St station and both utilities then wire the bank to send the cash straight over. The last two hundred will be along as soon as I pass Go again.

A weak lemon drink? Why, thank you nurse. Say, you'll never believe what I just bought.

Microsoft's Surface Duo phone hangs up, drops out of support

Re: A Pity

We daily bemoan the death of our Palm PDAs: not 'phone models, just simple PDAs.

The usability of the applications and ability to cram loads of data onto the tiny screens far exceeds the Android & iPhone apps we've seen[1] with more accurate pointing thanks to resistive touchscreens and stylus (about the only bit that can be bettered now, using a Wacom-style pen, but at much greater cost). Bliss they were.

And the use of Graffiti character input! Really easy to learn (less so in v2, thanks to HP lawyers), faster & way than touch keyboards (bleeping capacitive touch, just stops reacting, then switching layers to type simple HTML, why does the 'phone put DEL in a different place, keep on typing 'p' instead of DEL); you could take notes in a dark lecture room without watching the screen & without disturbing everyone else with the glow.

Sometimes think it is better never to get anything that really is functional snd could become a really useful day-to-day device because by the time using it is second nature, poof, it is gone and you'll never see a decent replacement.

Don't think too hard about modern gadgets or you'll lie awake in a cold sweat, fearing that they'll even take away decent keyboards and you'll never touch type again. Wait, what is this, a flimsy thin laptop keyboard? NOOOOoooooooo

[1] and you didn't need to scour the store for something usable, most everything was: there maybe Palm-quality apps lurking for Android, but getting to them through all the hype...

Elon Musk has beef with Bill Gates because he shorted Tesla stock, says biographer

Re: Tesla, the company doing the most to solve climate change

Even with the use of "minimalist"? What do those Acolytes think a flash car has? A spa? Two storeys and an Orangery? Antigravity?

Hey ho.

PS

/s on the Orangery - almost didn't, but I'm learning!

Re: Tesla, the company doing the most to solve climate change

Ok, ok, next time I'll up the sarcasm and add the clearly necessary /s tag.

Sorry to have misled you.

Re: Tesla, the company doing the most to solve climate change

PS

/s

Had hoped that the "minimalist" and "eschew frippery" were the clue, but sadly not. Hey ho.

Tesla, the company doing the most to solve climate change

As they are, and always have been, the only suppliers of an alternative to ICE vehicles, with their range of sensibly-priced, minimalist electric cars that eschew any kind of frippery for the sake of maximising the efficient use of the stored energy.

Billions of 'custobots' are coming online. Marketers may need to learn SEO for AI

Re: Why use BASIC when you can use AI?

Clever boy.

Scheibenreif nerd-sniped me into ranting about inappropriate use of tech and totally diverted my attention from his real concern, ensuring the continued fetishisation of "brands".

Why use BASIC when you can use AI?

> having AI turn a list of meals into a shopping list. Scheibenreif said that process could be informed by a human advising a personal model of their preferred brand of commodities like pasta, but that generative AI could reach its own conclusions about brands individuals might find appealing.

Seriously?

There were simple programs in BASIC for 80s home computers that turned your weekly menu choices into a shopping list. You could even do the inverse and get suggestions for what to make with what you have in the fridge right now. You do not need to waste power on any form of AI to do that task!

Especially not a generative AI - what are you expecting it to generate? A new and innovative way to make something that looks like stroganoff but is almost, but not quite, entirely unlike stroganoff?

Google thinks $20M ought to be enough to figure out how or if AI can be used responsibly

Re: Do No Evil

"Don't Be Evil" : Google, early noughties.

Tax evasion: Google, early naughties.

And nowadays

Google: Bard to the bone.

Linux 6.6's in-kernel SMB networking server graduates

Magical rusty thinking

> Unless this is formally proven or rewritten in a safer language...

(With the corollary that we all know there is only candidate for that "safer language" in the Linux kernel...)

Because, of course, the two are equivalent: if you write in Rust it is as good as a formal proof. Not.

Even if this guy was being hyperbolic[1] it is a bad idea to bandy around such an idea: there exist people who would love to take such an implied equivalence seriously. Without considering that, for example, memory errors of the sort Rust can deal with would be nothing compared to, say, an overly-privileged SMB server forgetting to honour any access rights[2].

Now, if there was a push from a section of the community to get code into the kernel that was written in a language amenable to formal proofs, and the accompanying proofs for each module, we could really start making some progress with a trustworthy OS[3][4]

[1] ref the mention of being paid in gold - which is a weird request but if he thinks the hassle of liquidating that is worth the hassle, each to their own

[2] no, not trying to say that KSMBD has this flaw, it is just an example of a logic/design/coding flaw that can exist without any bad memory accesses.

[3] you may say I'm a dreamer, and I'm probably the only one...

[4] btw, I'm nowhere near clever enough to create a proof of any program, let alone one worth running (I did the reading back in Uni, however...). Shameful, I know.

Microsoft to kill off third-party printer drivers in Windows

Amazing delivery from Reg article, as always.

> While some wags have dubbed the framework the "Unwanted Windows Platform", it's always good to see legacy tech being retired in favor of something with a bright future ahead of it.

How do you manage to type that stuff with a straight face? Bright future, oh goodness me.

So, if this horror comes to pass, to make full use of a printer's cleverest bits, I'll have to stop ripping out the whole Windows Store and all the UWP stuff? I suppose that might make the OS setup a little easier (a few clean ups to be stricken from the action list), but not looking forwards to having to live with the end result. When it gets to that stage, guess I'll not be bothering with any clever bits that require customisation to access <shrugs>.

Musk's mighty missile is ready for launch once FAA says OK

Re: "A lack of a sound suppression system"

> thata dust will go everywhere including being sucked into the rocket motors.

Rocket motors don't suck! See reply to your previous comment.

> The moon isnt hard rock with no dust, the dust gets everywhere...

Yes, yes it does. Unless you happen to have a large blowy thing that can push the dust AWAY - like, say, the exhaust gasses from a rocket!

Ok, once blown, it can settle onto stuff in the surrounding area - but it won't travel far, as there is no atmosphere or wind to carry it a long way (unlike on the Earth, where the UK is currently being covered in dust blown off a desert).

As far as risk to a Moon mission goes, the dust will be - as it was with Apollo, as you so correctly described - a nuisance when it is dragged indoors on spacesuits. It is very abrasive[1] and will oxidise (giving off that characteristic smell), so for long-term survival something will have to be done about it. Possibly something electrostatic in nature, as you don't want to be using water or any other consumable to clean it off (unlike after landing back on Earth where we can just turn a hose onto everything).

[1] abrasive because there is no wind on the Moon, unlike on the Earth, so the particles are not being continually rubbed against each other and having the roughness knocked off them - i.e. it is not being turned into sand. Compare the difference between volcanic "ash" and volcanic sand on Earth: the same basic stuff, only one has been weathered by the wind and the rain and the ocean and, well, the weather.

Re: Cost

> what happens that dust gets sucked into Starship rockets

*snigger* He thinks that rockets suck stuff in!

CowHorseFrog, you yourself have pointed out that the rockets push out GAS - they don't suck anything in from the outside! That is the entire point of a rocket engine!

Oooh, maybe this is why you referred to them as "aircraft" in an earlier comment?

Come on, at least *try* and learn a little bit about rockets - this is something we started covering in junior school, for pity's sake: quarter-filling a plastic bottle with water, pumping it up with air and Whooosh! I had a two-stage plastic water rocket when I was about 7 years old, nobody is trying to keep this sort of physics a secret from you.

Re: Cost

> So whats the point of having reusable rockets then if you cant land and reuse them from the moon ?

Because everyone, including engineering types who have been looking at this problem since, oooh, the 1940s, and everyone - except CowHorseFrog and a few total nutters, such as Flat Earthers - calculate that we CAN land them on the Moon and take off again - then land, take off, land, take off ...

Which is not to say that it *will* be StarShip that manages that - not because the whole concept is unworkable, as you claim, but simply because its development hasn't got to that point *quite* yet. Maybe next year.

Re: Cost

> which must have provided the protection equivalent to a launch pad.

Pure supposition on your part - you continue to claim that some protection is needed, so you glom onto anything that might support your conceit.

Try finding some actual evidence that the Apollo designers ever said "we can not take off without having a pad" instead of "there is no need to carry that lump back up into Lunar orbit, it has done its job and leaving it behind makes the rocket equation for the ascent so much nicer".

There is no shortage whatsoever of technical details from Apollo. You are making a positive claim, it is up to you to demonstrate the accuracy of that claim.

> those engines will be blasting shitloads of gas into a surface made of fine dust and lots of rocks, which must reasonably destroy the engines just like what happened a few weeks back in that launch.

Dust will be raised. Already admitted that. Now, think for a minute - what direction will that dust be going? Outwards and a little upwards. AWAY from the engines. Even you point out that the engines will be producing GAS - which will be moving outwards, carrying dust with it. Once the gas dissipates into the vacuum and no longer has the pressure to push the dust onwards, the dust will simple continue moving outwards, in a parabolic arc, until it lands on the ground again. How would it destroy the engines by moving away from them?

Damage on Earth was caused by the gas being restricted by the atmosphere - leading to lots of energy being carried UP by soundwave - and by dust and larger particles being supported by and carried along by the atmosphere.

The Moon does not have an atmosphere!

Re: "A lack of a sound suppression system"

> enough of the glasses hit the ground at a reasonable pressure

Ah, autocorrect.

But it is possible to generate thrust by throwing glasses out at high enough velocity (so take contact lenses as a backup, just in case there is a need for emergency extra power).

> When Apollo visited the moon, the lander was two parts, one part remained on the surface and acted as a pad protecting the return vehicle when it blasted back to the command module

Nah.

They left the bottom part because there was simply no point in taking it back.

Its supplies had been used up (the first landing was famously made with a rapidly emptying tank), they weren't going to pack up the rover (on later missions) and take that home.

> Your answer is all bullshit, you havent addressed how the blast from starship will be protected from the vehicle itself, because as we can see they couldnt do that on earth.

On Earth[1] there is this little thing called "atmosphere" (or "air" to use fewer syllables) which is thick and good at constraining the expansion of gas plumes[2] and transmitting the energy in that plume as "sound" (you may have heard of it).

> Big rockets make a big mess when they blast off,

And little rockets make a little mess, even on Earth with all its nasty gravity and horrible old air.

The Starship is a little rocket[2], the Falcon stack lifting Starship off from Earth is far away.

[1] not "earth", as you acknowledge they used a pad and not just sitting on the unprotected earth

[2] have a watch of any big launch - see how the plume of the exhaust gets wider and wider as the atmospheric pressure lessens, so it goes from a searing lance of light, heat and destruction to a rather pretty blue-tinged almost sphere that is many, many times wider than the vehicle itself.

[3] by comparison

Re: "A lack of a sound suppression system"

> . Gas not sound waves, GAS.

Yes, and? On the Moon that gas will immediately expand and be harmless very quickly. You do understand that the problem on Earth *really* is sound, don't you? The exhaust gasses push atmospheric gasses and they push other atmospheric gasses, carrying the energy of the exhaust gasses up and onto the rest of the vehicle structure and onto nearby structures? This transmission of energy occurs as pressure waves, which we call "sound".

> Are you seriously going to tell me an aircraft that large is going to be able to hit the moon

No, because it is not an "aircraft" - the hint is in the name! A Mars landing by Starship will also be in "rocket mode" for the landing portion, even if it manages a bit of aerobraking, so - ditto.

> is going to be able to hit the moon or mars or anywhere without throwing up a lot of rocks and dust ?

Well, the Moon, Mars and "anywhere" all have very different characteristics. Stick with the easiest, the Moon: yup, when it comes down on its 1/6th G thrust, as it gets very close (so enough of the glasses hit the ground at a reasonable pressure, instead of just harmlessly blossoming out the sides, dust will be kicked up. Maybe pebbles, doubt that "rocks" wiil be shifted. And it will all do a nice little (near as dammit) parabolic curve in the (still, despite that gas rushing away) near vacuum and come down a short distance away.

For the first landings, no trouble at all. When we land next to the 100-plus occupant colony, come down a kilometre away and use a bus.

> (lots of numbers copied from Wikipedia, absolutely none of which are telling you anything at all about the conditions when landing on the easy case, the Moon, let alone anything about Mars landings)

Well done, nice bit of cut'n'paste, but, if you don't mind me quoting one of your own earlier posts on this article:

> How about you actually address the question asked instead of talking about other things to divert attention.

Re: Cost

Let me have a think and get back to you, maybe the next time you ask this question.

Re: "A lack of a sound suppression system"

> How will starship work on the moon with no pad

You mean, just the top bit, not the entire Earth-launch stack? Just the teeny bit from the very tippy top? How will that ever manage without a pad?

> and no suppression system etc ?

Flame and sound suppression on the Moon? How *are* they going to stop the sound waves propagating across all that vacuum? Hmmmm....

PEBCAK problem transformed young techie into grizzled cynical sysadmin

Re: PICNIC > PEBCAK

> Be the "friendly face of IT".

Trust me, this _is_ the friendly face of IT.

Like young Wednesday, you don't want to see me when I'm smiling, as I enforce the company policy on your server directory.

You may even hear me give a friendly chuckle - although I am under strict instruction to hold that in until I'm alone in the server room, something to do with the PFY and PTSD.

AI to replace 2.4 million jobs in the US by 2030, many fewer than other forms of automation

Re: So we’ll be stuck at 2022 forever?

> we'll go continually backwards

The BBC showed us the result already: The History Of The World Backwards.

"set in a world where time flows forwards, but history flows backwards"

> It seems to me that practically everything the brain does can be reduced to pattern matching.

You need to allow for pattern creation, that is the interesting bit. Pattern matching goes on, but the patterns have (not yet) been determined to be intrinsic. Reference the experiments with development in an environment lacking in vertical or horizontal stimuli and the subsequent inability to recognise those stimuli later on.

Neural plasticity is not about pattern recognition, it is about (among other things, let us not be too simplistic) creating the ability to admit or prune signals based upon apparent significance and then set up means to recognise patterns in those signals

It is a hard and unsung life, being an AI researcher.

They take a problem, one that hasn't been solved before by anything other than a human being (and even then, not by all humans). As soon as they find a way to emulate the behaviour, then publish how they did it, everyone is all "well, *that* is obvious, isn't; call this AI research, all you've done is just...".

That is just - an "obvious" way to prune search trees, a statistics prediction method, a pattern matching exercise (*and* you threw away most of the image to do it").

If it can be made to work, then it isn't an AI method any more, it is just common sense; you failed, try again (or not, depending upon how vitriolic the person feels).

[Their] strategy should include investments, guardrails, and checkpoints

> the report concluded

Amazing.

Such a novel and well-reasoned course of action, totally unlike the way any company has ever responded to anything else.

This conclusion alone means that the report is worth reading. How they manage to think this stuff up just boggles the mind.

Guy who ran Bitcoins4Less tells Feds he had less than zero laundering protections

Suspicious packaging

> the cash was often packaged in a suspicious manner, including inside hidden children's books, concealed inside fake birthday or holiday presents, buried within puzzle pieces, or wrapped within multiple magazines

Those sound like all the ways I'd stuff a £20 or £50 (for special days) into an otherwise cheap and grotty gift for the youngsters when I had no idea what they really wanted for Christmas.

Maybe that explains the vans - we all know the drain is further down the road...

d-Matrix bets on in-memory compute to undercut Nvidia in smaller AI models

Fairly short window of opportunity

LLMs are very much flavour of the day, in much the same way as everyone talks about the latest fantastic bridge project (the press around the building of the Clifton Suspension Bridge was vast and colourful[1]). Meanwhile, there are many more people are being regularly served by all the hundreds (thousands?) of smaller footbridges and unglamourous overpasses put up in the same timescale. Similarly, for each honking great LLM there are numerically many more tasks that can be tackled with smaller ML models.[2]

If d-Matrix can avoid getting sucked totally into the vision of selling increasingly large systems to make a small number of big sales, before being overrun by the likes of nVidia, there should be a healthy long-term market for decent model-running chips that can be used in smaller products, that sell in larger numbers. Some alternatives/competition down there would be a Good Thing, especially if it can bring down prices and become well-known outside of specialist circles. Okay, they are still up against nVidia (Jetson) but that is put at the pricey end, leaving much room below where there is, what, the K210, anything else?[3]

[1] actually, mostly b&w 'cos the newspapers back then - look, figure of speech, ok!

[2] not really an amazing observation: so long as it is possible to build a smaller working X you will see more mini-X than maxi-X in use: consider JCBs and the cute little mini-diggers.

[3] genuine question, do you know of other devices in that range or lower that help run your smaller ML models? Actually available as hardware, not just drop-in IP for your next fab run?

Google rebrands 'android' as 'Android' to remove any doubt about its affiliations

> The black android icon in the bottom right looks suspiciously very much like the top of a classic dalek.

Came here to make the same observation.

But I was wondering what was missing, then realised: of course, it would confuse the branding to give it an iStalk.

The Anti Defamation League is Musk's latest excuse for Twitter's tanking ad revenue

Re: Lesseee here.

Unless he was referring to Elon's grip on reality.

Re: Musk's Master Plan

Fingers crossed, that is what the AC was doing.

But that was a risky strategy, going without the /s tag.

Freecycle gives users the gift of a security breach notice

Re: MD5 Hashed Passwords

Most of the other sites reporting this are claiming that the hash used was MD5, but there is a distinct lack of corroborating evidence (like, say, a statement from Freecycle).

For example,Tom's Guide is citing Bleeping Computer, and they are citing: "The stolen information includes usernames, User IDs, email addresses, and MD5-hashed passwords, with no other information exposed, according to Freecycle."

BUT that last link, the one to Freecycle? There is no mention of MD5 in there. Hmm.

Other sites are just repeating this, generally without even bothering to give a citation at all: TechRadar gives no direct citation for MD5, ditto GridinSoft, ditto SecureBlink.

From all this, the fact that The Register doesn't claim it was MD5 can be attributed to their being the only ones who felt like _checking_!

There may actually be a genuine reason to believe it was using MD5, but none of the above could be bothered to provide a citation and I can't be bothered to check any more "news" websites or "professional company" blogs - it is all too, too depressing.

AI coding is 'inescapable' and here to stay, says GitLab

What measure is over-confidence?

> The report concludes by noting that respondents with more AI experience were less likely to associate AI with productivity gains and faster cycle times.

So we are left with a large pool of people without "AI experience"[1] who are being naively optimistic during the development phases (including code review etc as outlined in the article).

I was going to make snide comments about entrusting testing decisions to the AI, then, hey, at least they are doing *some* testing!

Say, are there are any stats about how long after deployment the major CVEs start to appear? Are we expecting a few house of cards to come falling down in 6 months? A year? Two?

[1] and, btw, no solid definition of what "AI experience" means - someone who has run Stable Diffusion at home? Or has followed some tutorials on training a simple Neural Net? Or has a research degree in AI from 2020 which only really covered NNs? Or has been involved since the 1980s and has a broad grasp on different AI models?

The world seems so loopy. But at least someone's written a memory-safe sudo in Rust

Re: NP complete

> the cyclic references described won't be GC'ed

Great strawman skills, deliberately only quoting half the sentence. Try responding to this quote:

>> the cyclic references described won't be GC'ed because there is still at least one active reference to that structure.

Yes, clever boy, GC will collect cyclic structures. That has been the case since the 1960s.

BUT so long as you have an ACTIVE reference, the GC "mark" phase will, correctly, NOT set that structure up to be reaped by the "collect" phase.

In fact, you even describe various *forms* of active reference (weak, strong)!

Re: NP complete

Bad phrasing on my part: the cyclic references described won't be GC'ed because there is still at least one active reference to that structure (the one being used to scan what was intended to be a list).

The actual problem being highlit is that the program is in an infinite loop, counting the entries, because creating - incorrectly - the cycle wasn't prevented by GC (as was implied by the post I replied to - i.e. the claim that GC solves all memory errors).

Microsoft billing 3 cents a minute to revisit tedious Teams meetings via API

So they are already paying for a service, then won't this save them a good chunk of that?

So not something to balk at.

Haven't seen that Netflix example (may try later) but the near-as-makes-no-odds real-time transcription at last year's EMFCamp was impressive. I assumed (ok, yes, never do that) that was machine transcription, given the nature of that camp.

But even if you have a few boners in the transcript, at that price it is still good value.

> That'll start to add up if they want recordings/transcriptions of every meeting for regulatory recording keeping purposes.

These are brand new regs, are they?

Otherwise, how are they complying already at a rate of 5.4 cents a minute (assuming you get charged to record in order to transcribe)? That sounds like an absolute bargain for transcriptions![1]

[1] random "first page I found": The average price charged by transcription businesses in the UK for standard two speaker interviews, conversations or discussions is about 90p to £1.50 per minute, although for a traditional transcription business this price will very much depend on the turnaround time. Recordings for 3 or more speakers can be considerably higher as they are much harder to transcribe.

Meeting price ticker

Idea for app: put up a price ticker on the meeting room big screen, which shows the current MS Teams cost so far - and the name of the manager whose budget it is coming from.

Oh, and a variant that shows the total being charged to the company account for *all* the meetings in progress and monthly total. For use by the CFO.

Goal: reduce the length of meetings[1] and increase employee productivity.

[1] oh, sometimes I really crack myself up; as if anything could ever shorten meetings! Ah me. They'll be taking it out of the biscuit fund instead.

Guild behind actors' strike fears video game workers also at risk from AI

Big difference is that CGI came on-stream very, very slowly over decades: we got some really ropey stuff that was only used because it was still better than using a black backdrop with holes for starfields. As it got better, everyone could adjust. It helped that a lot of it was laughable - e.g. the animated acrobatics in that daft Van Helsing movie.

Plus acting against a real set still has advantages of quality, knowing where everything is without having to keep it all imagined - or just not corpsing when trying to keep an eyeline with a tennis ball on a stick (blooper reels can be an education in themselves!).

Actors have adjusted to the use of Serkis Folk and seeing themselves replaced by CGI, but they still did the acting.

But this step is a short, sharp, shock as they get snipped out of the process entirely.

Farewell WordPad, we hardly knew ye

Would sorely miss WordPad

The only "word processing"[1] I need to do nowadays (thankfully) is the occasional one or two page letter for hard copy - WordPad allows that to be written easily & without having to wait for LibreOffice to load up (well, that is what it feels like). And it is a lot less cluttered (I don't need to use all that WordPad can do, let alone LO!).

Very glad to have seen the URL to the MS github, above; hope I don't forget to snarf a copy of that!

Still need to keep LO around for all those folks who send out weird format files in huge emails...

[1] proper big documents - like documentation - get written in plain-text with markup, pick whichever is suitable (Doxygen, Markdown, LaTeX, weird home-grown macro language thing), kept under version control and run off as a PDF if anyone wants a collated printable copy.

Right to repair advocates have a new opponent: Scientologists

Re: Expose,

Can anyone spare a closing quote?

Re: Just another variation on the useless "lie detector"

Given how the meter needle is deliberately built without any damping mechanism at all, it will react to any old noise strong enough[1] to be picked up - which is one of the classical[2] ways of making an RNG.

[1] funny how everyone in the Sea Org is suddenly "unclean"[3] whenever the diesel generator kicks in

[2] now begins the discussion about how biassed such an RNG will be, it isn't _really_ random, not good enough to be included within /dev/rnd...

[3] unclean, unthetan, unclear, unmutual - who knows.

Re: Expose

Isn't the book of "Battlefield Earth" the one where they get a few bombs sitting together on a teleporation pad inside the aliens' home and then say something along the lines of "we'll explode the first one and if that doesn't do the trick, we'll try the next two together"?