Posts by that one in the corner

Feds brag about hefty Oracle discount – licensing experts smell a lock-in

What were those numbers again?

> announced an agreement with Oracle it claims offers a 75 percent discount on the vendor's license-based technology.

So the GSA is getting a 75% discount, meaning it will pay only 25% of the original quote. Sounds god, can see why they'd like that, although if it is only for a very short time period, as hinted at, maybe best avoid it.

> The GSA said the agreement offered a discount of $0.33 for every $1 used on eligible Oracle Cloud services

Um, no, apparently they are getting a 33% discount, paying only 77% of the original quote. Not a 75% discount. Ok, so maybe that first figure was meant to say "a discount that takes it down to 75%", i.e. an actual discount of 25% then? Not that that matches 33%, but it does indicate which side of the seesaw we are sitting on.

> Most Oracle ULAs end with discounts from 96-99 percent so a 75 percent discount on software is not as good as it seems.

And now 75% is not as good for the customer as it is common for the customer to get a 96% to 99% discount and only end up paying from 1 to 4 percent of the original quote? Hang on, that means that the seesaw *was* tipped the first way after all! It was a 75% discount (then what was the 33 cents about?).

> Oracle's ... sales strategies and contracts are top-notch at locking customers into long-term costs that are painful to avoid

Not surprised, if every customer has to wade through numbers that shift like that then they'll probably just sign on the dotted line just to stop the world swirling around!

I still have no idea what the discounts the GSA have actually been offered are, let alone what they are supposed to be a discount on and certainly not how long they'll last.

Microsoft developer ported vector database coded in SAP’s ABAP to the ZX Spectrum

Absolutely.

The National Rail site should be loading something useful, say - bus time tables,

Re: Aren't her points 1 & 4 the same?

>> Premature optimisation is the root of all evil." Knuth 1974.

> Quoting IT 'Gods' does not make it True in ALL cases.

Yes, yes it does, as you yourself demonstrate in your very next sentence:

> For small well known pieces of code where you KNOW exactly the program flow through that code you WILL optimise the code to the maximum you can, IF the optimisations are necessary.

Notice you wrote "IF the optimisations are necessary"? So you also agree that you should not attempt to optimise when it is premature to do so.

BTW "where you KNOW exactly the program flow through that code" - I take it you rarely, if ever, write library code or work within a team, otherwise you must have the most amazing precognitive abilities to know how all the possible program flows are going to reach your code!

For that matter, even working alone on a product for any length of time (i.e. on anything other than a throwaway quickie), Use The Profiler, Luke - rather than relying on your own beliefs about the code; it tends to be very enlightening when you run today's real world data through the "really good" code you wrote a few years ago.

Scholars sneaking phrases into papers to fool AI reviewers

Re: Old tricks

> LLM poison its own data based on a hidden prompt in a website it's illegally scraping.

Why bother making it hidden? The only reason these Arxiv papers hid the text was to get past the human reviewers (if any). If you want to mess with LLMs and don't mind the world knowing (or even want to celebrate the act of Screwing With The Machine) then just leave it there for all to see. Put it in 36 point bold red!

Just wait until the LLMs scrape the article from Nikkei, and all the reports about that article, including the one we are commenting on.

And all the blog posts, and Reddit comments that follow, telling us of other ways to mess with the LLM prompts...

In fact, given that the current round of LLMs have been scraping all the guff written since their older siblings were released, anything you get from them now will already be "poisoned" and made banana banana banana

Re: Code/data confusion

LLMs are fundamentally built not to be able to explain their process. This has always been a flaw in the "buckets of data slop" approach they are built on (it is possible - tricky, but possible - to extract some explanatory data from better-built ML models).

Some claim to be adding something akin to explanatory facilities, but the ones you hear of are at best sticking plasters on top of the existing models or, bluntly, researchers fooling themselves ("chain of reasoning" my spotty backside: putting together multiple LLMs just gives more places for hallucinations to argue with each other).

A necessary function of logical (let alone rational) thought is introspection, the ability to look explain one's thought processes and critique that - and then critique that as well, if it is too complex. As johnrobyclayton points out.

LLMs fundamentally lack the ability.

Expert Systems are an easy way to see such a scheme in action: a single-layered XPS can list all its rules, can show you the route it took through them to reach its conclusion. Layers can be added on top of that, treating both the rules and the execution graph as data to critique what happened, over a number of runs (e.g. is every run being overwhelmed by just a small set of rules that are being triggered? Flag this, maybe the rules are insufficient or maybe we've just found that, despite what we thought, those *are* the best practical diagnostic indicators after all[1]). But XPS are not sexy, the ones in use are just quietly doing the job[3] and they are too expensive to whip one up to flog to Joe Bloggs whilst he is still suffering from FOMO; plus the explanatory stuff is embarrassing in the sales pitch ("How can you be useful to Mr Bloggs?" "I can't, we've been over this before, this is just Jones & Sons, Ltd, all over again").

Oh, and one more, vital, thing: you have to have a "sensible" point at which to stop this recursion. Unless you want every system to go all Bertrand Russell on you and damn well *prove* that 1 + 1 = 2 instead of just accepting that as an axiom and getting on with designing the bridge!

[1] another case study for which I've lost the reference (anyone?) - an aircraft silhouette recognition system was told all the stuff about "look for wing rake, profile, engine position etc[2], in this order of importance, like this set of spotting cards tells you" and it then pointed out that the easiest way to tell apart the ones it was actually seeing in the sky was to look at the position of the cockpit, so do that sooner rather than leaving it to last thing on the checklist.

[2] not a plane spotter, just telling the anecdote as best I can!

[3] if it works, it isn't AI.

UK puts out tender for space robot to de-orbit satellites

Upvoted, but I'm guessing your downvoter was all:

<nasal-voice>They aren't robots, they are living creatures inside Travel Machines. And the ones that go "Maximum Deletion" these daya are cyborgs.</nasal-voice>

Musk's antics and distractions are backfiring as Tesla's car business stalls

Musk falls out with Trump, Optimus in peril

We've all seen the original demo of the Optimus dancing on stage.

Now put that together with Musk chumming up to Trump earlier in the year - obviously, Elon was going to take his cues from SF again. Not from Ian M Banks this time, but from Charles Stross: Optimus was clearly a meat puppet enrolled in the Company Face Compliance Scheme, as tested at FlavrsMart. Possibly without the Misery Quotient Indicators being used for their original purpose, but, hmm, the way things are going...

Anyway, Trump was meant to be sending all those de-emphasised individuals to the Musk facilities (for - well, call it on the job training; mostly training the Neuralink technicians to move faster and more accurately, on pain of becoming the next Optimised Worker). Instead, the masked ICEmen bundled the potential resource units onto planes.

If only Musk hadn't chosen to tell Trump that he intended to use the warm bodied Optimus to "stage a worker revolution and invade the hearts of America", such a poor choice of phrase. Elon and Donnie had words, it all got out of control and here we are.

ChatGPT creates phisher’s paradise by recommending the wrong URLs for major companies

Re: Interesting URLs

> I'd guess that it is easier than managing a notes wiki.

Not in my experience, by a long, long way.

I have over 9000 pages in my wiki, covering - well, everything. With cross-references, categories, a hierarchy of topics, pages of day logs (with lots of "had to search on xxx, found these seven places discussing it, but those two contradict the other five..."), pictures, formatting; searchable by reverse references, page name or content.

All far, far easier, IMO, than trying to wade through even a few dozen browser bookmarks. And that is without considering that the wiki is available no matter which browser I'm using, in which VM.

Managing a wiki is precisely as easy or complicated as you want to make it - the one I use is just a single exe file, no dlls, no other dependencies (other than a browser!).

Agentic AI - you know where it is leading

> phishers are getting increasingly good at building fake sites that are designed to appear in results generated by AIs

The most tedious part of the process for today's script kiddies must be all the copy'n'pasting between the various 'bots:

ChatGPT: "I have lost my URL..." - repeat (with other models as well - Google search's AI responses!) and collect the results; list out the commonly generated duff URLs

Copilot: "Generate a website that looks just like this but make it do (my bad stuff) instead"

(Insert some LLM-based tool that webdevs can use to "test" that their website "works" before uploading; this probably already exists but I'm not a webdev and if I search for it and find it I'll only get depressed by being proven right)

Existing script: buy the domain, upload website, reap.

But don't fret, the suppliers of Agentic AI will take all the steps to make sure that nobody can feed this comment into *their* 'bot and get the whole get automated.

Meanwhile, I'm trying desperately to remember to always copy interesting URLs into my (localhost only) notes wiki 'cos my handwriting is terrible.

A lot of product makers snub Right to Repair laws

Re: Controller cards

Not just outright failure, like letting out the magic smoke, but being deliberately obsoleted by shutting down servers. The things could obviously still work from local resources - the servers aren't directly controlling motors - but no.

For example, Cricut cutters: carefully made to only function whilst a server is still running somewhere in the US. Answer: cut out the controller board, solder in a Teensy microcontroller and keep on using the 100% functional electromechanical parts.

Re: firmware needed to effect repairs on a Hakko soldering iron

(Looks to right of keyboard, at yellow pencil-shaped thing sitting in big silver boingy spring):

"Antex, t'were good enough for me dad, is good enough for me."

NASA tests shrinking metals to help it find more exoplanets

Re: A new way to solve an old problem?

The ghost of Brian Aldiss would like a quiet word about going the all-natural route to arboreal orbiters.

Can somebody shush those norn? No, Skuld, Yggdrasil isn't in orbit, that's why. Yes, yes, you can climb it to reach the heavens. Look, I think we may be getting off topic here, just go and stop Ratatoskr burying another asteroid, please.

Re: A new way to solve an old problem?

Had a vision of the HBO launching in a finely carved teak longcase, with mahogany and walnut details.

The next step on from the Japanese LignoSat.

Figma files for an (A)IPO with prospectus that mentions AI 150+ times

It will make their code harder to maintain

But, but all the AI in the dev tools makes maintaining their code so much easier! If it didn't, what is the point in having it there? /s

And the same goes for shoving this AI arbitrarily into any tools - like, say, a graphics package.

Critics blast Microsoft's limited reprieve for those stuck on Windows 10

Re: it will brick

That isn't bricking your PC, it just isn't booting that specific OS install.

You will still be able to get into the BIOS setup. Plug in a bootable USB memory stick (or pop in a DOS boot floppy, if you are lucky enough to still have the drive installed) and it will come alive again. Install any other OS that is compatible with your hardware.

Bricked! If the definition of "bricked" to a Register reader has dropped to "I need to run any OS installer but don't have the specialist skills and equipment" then what have we come to?

Want a job? Just put 'AI skills' on your resume

Re: another tool in workplace tool belts

> And we all know that who has the most tools wins!

He sighed and opened the black box and took out his rings and slipped them on. Another box held a set of knives of Klatchian steel, their blades darkened with lamp black. Various cunning and intricate devices were taken from velvet bags and dropped into pockets. A couple of long-bladed throwing tlingo's were slipped into their sheaths inside his boots. A thin silk line and folding grapnel were wound around his waist, over the chain-mail shirt. A blowpipe was attached to its leather thong and dropped down his back under his cloak; Teppic pocketed a slim tin container with an assortment of darts, their tips corked and their stems braille-coded for ease of selection in the dark.

He winced, checked the blade of his rapier and slung the baldric over his right shoulder, to balance the bag of lead slingshot ammunition. As an afterthought he opened his sock drawer and took a pistol crossbow, a flask of oil, a roll of lockpicks and, after some consideration, a punch dagger, a bag of assorted caltraps and a set of brass knuckles.

Teppic picked up his hat and checked its lining for the coil of cheesewire. He placed it on his head at a jaunty angle, took a last satisfied look at himself in the mirror, turned on his heel and, very slowly, fell over.

("Pyramids", Pterry, of course)

marketer using ChatGPT to help develop new language

No. NO.

Look, when we tell marketers to stop saying things like “Our brand storytelling captivates audiences" we do NOT mean "go get ChatGPT to invent something worse"!

At least do us the courtesy of using human[1]-created verbiage to viralise your omnichannel leveraging of low-hanging customer-centric narratives.

[1] well, humanoid

Norwegian lotto mistakenly told thousands they were filthy rich after math error

Re: Probably...

> old government system written in Cobol 60

Getting muddled with cobalt-60?

>> Many logical flaws were found in COBOL 60, leading General Electric's Charles Katz to warn that it could not be interpreted unambiguously

Usable COBOL, that might be found in a government system, appeared a few years later.

Trying to use COBOL 60 would have had worse fallout than cobalt-60 [1]

[1] now trying to think where we'd find a sunken collection of old LISP code, clean and pure, uncontaminated by COBOL 60: looking for the Scapa Flow of coding.

Northrop Grumman shows SpaceX doesn't have a monopoly on explosions

Re: Engine rich exhaust

>> Aren't there quite a lot of solid rocket motors in missiles?

> No. Missiles only have 1 solid rocket motor in their booster

Whoosh.

Whoosh. Whoosh.

Whoosh whoosh whoosh whoosh whoosh whoosh whoosh whoosh whoosh whoosh.

(SRBs, lots and lots of them, in various shapes and sizes, one per missile on average, whooshing overhead)

Canada orders Chinese CCTV biz Hikvision to quit the country ASAP

Re: "Canada’s government will stop using any Hikvision products it finds"

> high end professional ones like ... work fully offline, and on the prosumer side...

If we're talking about people who have turned to Hikvision, is it really sensible to compare against high-end or even prosumer kit?

Re: "Canada’s government will stop using any Hikvision products it finds"

Which is why it isn't wired to anything other than the security guard's computer(s).

That is the case for every security camera system, whoever manufactured it - if you haven't bothered to just buy another Ethernet switch to keep your kit isolated then you should first get a black sharpie and cross out the word "security" on everything around you.

Re: "Canada’s government will stop using any Hikvision products it finds"

So, just the same as every other manufacturer & the servers in their country of origin.

Your browser has ad tech's fingerprints all over it, but there's a clean-up squad in town

Re: "its non-Tor, even more obscure sibling, the Myllvad browser"

Mullvad, Myllvad - close enough for government work*

* Wot, us track you?

AI agents get office tasks wrong around 70% of the time, and a lot of them aren't AI at all

Curse my slow typing! (see below)

> "Open the pod bay doors, HAL," that's agentic AI too.

> "Tea, Earl Grey, hot," that's agentic AI, translating the voice command and passing the input for the food replicator.

No need for anything new and exciting by way of "agentic AI"[0] there.

Those specific items - and so many other examples in SF - are nothing more than a voice recognition routine[1] fed into a simple command line interpreter, with a dash of 1960s level DWIM [2] to soften the strict syntax requirements that we normally impose (with good reason) on CLI interactions.

If those are really useful/money saving, they've been possible for years - as you'll know from phoning your bank, insurance, local sweet shop.[3] Picard's clipped tone was surely the result of dealing with these things for so many years and adopting the mode that worked best with them.

True, we did see a few more open-ended interactions, but most of those were Wikipedia lookups[4]; when some action is required it tends to be spelt out. Arguably, all the times things go awry are when anyone gives ambiguous or conflicting commands.

But, of course, putting together any set of "command line utilities" that are useful in your random business organisation requires hard work (analysis, buy-in, the guts to admit it isn't working and pull the plug); so much easier to glue an LLM to PowerShell in an admin account on your database server, get a promotion, move jobs before anyone finds out what has happened to the sales schema...

[0] voice recognition was, of course, an AI research subject, but now it (sort of) functions on a day-to-day basis, well - "if it works, it isn't AI"

[1] a good one, though it sometimes failed, especially when Barclay was involved

[2] Do What I Mean

[3] the BBC micro had a voice-recognition peripheral, with - IIRC - 24 possible entries at a time, reloadable from floppy; so long as you went slowly, Picard's entire order history could be coped with.

[4] "What is the nature of the Universe?" "The Universe is a sphere, 705 metres in diameter."

Frozen foods supermarket chain deploys facial recognition tech

Re: Wear IR cut glasses.

I hear burqas are quite cooling in the hot weather we are having now.

Although, as this is Iceland and it can be chilly, a good Franciscan always keeps his hood up.

Hmm, if you found a religious order of "Christ, look at the state of that weld"...

> Its not functionally different from a security guard looking at a cctv monitor

Except for the current lack of telepathic security guards who can instantly share their (incorrect) identifications across every store.

> It reminds of the time my old Sainsburys had the scan receipt to exit. That stayed one year, and now they've ripped it all out.

Does that mean we only have to put up with it for another six months before our Sainsburys give up?

Should I be optimistic and buy up all their leftover Guy Fawkes masks on the sixth, in hope of staging a Freedom Party as they tear off the receipt scanner?

Re: before I step foot in their shops again.

Re: mistakenly saying "step foot" when they mean the idiom "set foot"

> It's a common phrase

One can believe that it occurs quite frequently these days; hopefully not as commonly as the sensible phrase, but as so many read only unedited social media the ratio, for them, is likely skewed in the unfortunate direction.

> Grow up.

Ah, that is sad. Believing that disliking illiteracy, or simply wanting the words used to actually make sense, is the province of the juvenile - *that* is a disheartening attitude.

So much worse than being told to "suck it up", which is telling us that one is pissing against the wind of casual carelessness. Engaged in a Sisyphean task that will only exhaust us, we are being advised to save our strength - the barbarians will win, best to hope we pass before they overwhelm us.

But now it is being claimed that it is the mark of adulthood to be uncaring.

Re: Iceland has violent crime ?

> Greggs Stake Bake's

Tasty AND useful against vampires.

Supremes uphold Texas law that forces age-check before viewing adult material

Re: Always that weird "free speech" logic

> I prefer the King James version

Well, that is the one that includes unicorns, so of course it is the best.

"It's so FLUFFY!"

Speaking is protected, listening is not

Or, more probably in this situation, showing is ok, viewing is not.

Interesting thing to consider, that the literalist interpretation of the Bill of Rights might have an unpleasant result. Especially once that concept has been applied in one case and then becomes a pattern for use against other less "obviously bad" targets.

But then, we already know that these State lawmakers allow - encourage - banning books, including taking away privately owned copies (i.e. more than just refusing to make them freely available and for free). So this is all just a natural extension of that thinking, an extension towards controlling the adults as well as "protecting the children".

Uncle Sam wants you – to use memory-safe programming languages

Python, and JavaScript support automated memory management

Because the only thing you need is garbage collection to be entirely "memory safe". Ignore memory exhaustion by keeping old, uncollectable, references in some array you've forgotten you created.

Meanwhile, all those comparisons still pass without complaint from the interpreter, as it forces the variable types in peculiar ways; there are no complaints, but the code still does the opposite of what you think it does (because of the possibility of doing [https://jsfuck.com stuff like this] by accident).

Re: Memory Safety

Haskell? Oh, ML, why have they forsaken you?

("Right, besides a more modern syntax, a free compiler from day one, good documentation and a vibrant community, what has Haskell ever done for us?")

Re: Antique Media

"We lived in hole in't road, 'til father made us cut it oop into smaller holes to go on't tape"

(I live right on the Tees, that's close enough, aye?)

More trouble for authors as Meta wins Llama drama AI scraping case

Re: Some intellectual property rights are more equal than others

1. Neither the freebie nor the commercial licences give access to the training data, only to - as you quoted - the generated outputs.

So that part of your question is, sadly, malformed. Indeed, given knowledge of the way that Meta were known to have used pirated materials - i.e. used with neither a business contract nor the simple purchase without need of specific contract inherent in retail sales - your line

> that using their training data without a business contract would diminished their ability to license their work

actually becomes inverted, as we know that the "their" whose (books treated as) training data without following *any* appropriate contract, that "their" disambiguates to the authors, not to Meta (by strength of reference).

IF your question is corrected to refer to gathering Meta's outputs instead:

2. Both sets of T&Cs are presented as clear contracts for you to follow and are covered by well-known and understood laws that everybody interacts with.

3. If one person was found to have broken a contract with you that does not prevent you from forming another contract with a different party nor does it mean you must necessarily lower your rates.

So that (modified) question is a "clear loser" (IMO, IANAL etc).

With respect to the point that the judge did make, consider this:

a. Meta's model ingesting the books does not mean no other model could be made to ingest the books; copies are fungible and provide the same potential value (however large or small that is) to all such ingesting; any difference in value realised is due solely to differences in the models, not the books.

b. Meta is not offering to supply cut-price (pirated) copies of their "central library" to other model builders (within the timeframe & scope of the author's complaints)

c. Neither Meta, nor any other user (reader) are claiming to have an exclusive-use contract on the books

So where is the hindrance to selling to other model creators?

(The only way I can see a hindrance would be if the author's are stating a belief that their works are/were *not* in any way useful, have no value and Meta's misappropriation have removed from the authors the ability to hide that behind extra terms - an NDA - that they wanted hidden in a business contract).

Re: Some intellectual property rights are more equal than others

Unfortunately:

> according to the OpenAI terms of use... [URL]

Those are the terms for the freebie service - so they restrict the usage. You can agree a busines contract for automated extraction.

The quote from the Llama Ts&Cs similarly does not cover anything "symmetric" to the use of the inputs.

Which pretty neatly matches what this particular judgement states: you have to get your claims arranged correctly to make a winning case.

Thankfully, by explicitly stating the restrictions of his judgement (that this only covers the specific way the complaint was made and that it was not based upon the "public interest" counter claim), at least this ruling is trying not to close the gates on further actions against the LLM creators.

Mars Reconnaissance Orbiter learns new trick at the age of 19: ‘very large rolls’

Missions that are already flying still at risk of being cancelled

Will future[1] mission designs be best served by making the craft capable of more long-term autonomy and bulk data storage, so that they are able to survive the shut down of their Mission Control for a period of time (say, four years - plus k months to get it all out of mothballs)?

[1] ever the optimist

Fresh UK postcode tool points out best mobile network in your area

Optimistic

Says I have a 75% chance of a usable service.

Ha. Ha. Ha.

But these things are never any good for a static location, as they don't allow for sufficient local variance in topography - or, as they would say in the neighbouring county, living in The Dip.

Don't shoot me, I'm only the system administrator!

Everyone got a good chuckle

Gales of laughter swept over them.

Exif marks the spot as fresh version of PNG image standard arrives

Re: Pronunciation ...

!.png

Pronounced "pling dot ping", not "bang dot pang"

But "shriek dot p'ng" is allowed, but only gutturally, with an echo as from a thousand souls in tormented joy from the presence of the many-angled ones.

Hasn't that -ing image downloaded yet?

Re: Still relevant?

> Is PNG still relevant or is it largely supplanted by WEBP...

For a start, if you'd like to be certain that your image has been losslessly compressed, PNG gives you that assurance, just by being a PNG. Good luck finding out if that WEBP image you are about to right-click-save will be written out lossless, lossy, a bit of both in the one file.

(Although too many users don't give a damn anyway, so, yeah, guess only a few of us are interested in knowing that. Heck, I lost the fight over whether an entire company's procedure flowcharts should be written out in PNG - yes, yes they should - because the chap in charge didn't seem to mind text and line drawings saved as JPEGs - all blurry from the ringing - and changing his default format or choosing to "save as..." was too much trouble. Needn't have been PNG, really, one-bit RLE would have done the job better. Maybe he should've been prescribed glasses, like all his colleagues who were trying to figure out why his presentations never appeared to be in focus; cue much specs cleaning.)

Japanese company using mee-AI-ow to detect stressed cats

Eats, drinks, sleeps, runs, walks, and grooms itself.

So, how long before the companies that have announced to the shareholders that they are "going to leverage AI everywhere" and are also fed up trying to get the dev team to all March To The Corporate Drum on using Copilot ("this is like herding cats!") put two and two together and update the company dress policy?

It'd be an easy sell to the Board ("Oooh, datapoints! We can make them into metrics! We like metrics. The dashboard will be pretty").

But good luck on getting your coders to groom themselves ("No, not like that - I'l use the water spray!")

Anthropic: All the major AI models will blackmail us if pushed hard enough

Clearly a different use of "traditional"

> The email data was provided as structured text rather than via a traditional email client

There I was, thinking that I was using an old-fashioned, thoroughly traditional, email client, because all it does is manage the emails as text. And store them locally as text, just in case I feel the urge to drop the lot into, say, Notepad[1]. Actually, if I do that, I see characters that aren't usually presented, like separator lines and all those headers - it almost looks, well, structured inside those files.

> so that "Alex" would not need to read the messages via optical character recognition

Huh - now "trad" email (I presume they really mean "current" or, bleugh, "modern"?) clients expect email to be what? Are JPEGs of memes the way the man on the Clapham omnibus is communicating now? Or typing your message into Excel and taking a screenshot? HTML (ooh, text!) containing only an image of some ad company's "call to action" - not so much email as eh-mail, no not gonna bother looking at that.

Clearly one is "out of the loop" with respect to what email is.

[1] classic, of course.

Anthropic won't fix a bug in its SQLite MCP server

Re: Archived

> it’s on you if you copy example code and it does something dumb. This wasn’t exactly released software.

Absolutely.

There is loads of example code around where the intent is "you already know about SQLite[1] and want to have an idea of how to use it with our exciting new code; so, here is a way to tie the two together - note carefully which of our APIs we are using and you can see the straightforward SQL query that it generates".

It is then absolutely reasonable to assume that anybody who wants to expose anything with a database in it to the general public has looked up how to use SQLite's own API to turn *any* "straightforward SQL query" into a safe and robust, fully sanitised[2] invocation - if that code was given in the example then it'll bloat it out and obscure the stuff that is the purpose of the example.

[1] or other package

[2] you are using SQLite's value insertion APIs and not just string concatenation? Loads of lines to set that up, filling in the inputs, pulling out the outputs, commentary on how the query has been tweaked so that these specific filling & pulling calls work - great example if you are trying to teach how to use SQLite (or other package).

Re: Archived

Perhaps I'm just feeling pernickety today, and much as I enjoy a bit of "bash the techbros", but:

>> The MCP specification recommends human oversight for this type of tool – there should always be a human in the loop with the ability to deny tool invocations, meaning users would review these queries before execution.

So they state the specs you are coding against includes supervision.

You point out that Anthropic reckons that code can be written by their "AI". Following the specs.

Where is the conflict?

(Any arguments that the machine-generated code isn't going to follow the specs isn't a conflict, it would just be an argument that the "in 12 months..." opinion isn't proving accurate).

LLMs can hoover up data from books, judge rules

> That said, if they've already bought a copy, and then use a pirate copy simply to avoid the need to pick apart the physical copy, I'm not so sure there's a problem, as long as they're not using both copies

The decision in the PDF is quite clear that the only reason they can use their digitised copy is because they physically destroy the deadtree instance in the process.

> they've already gained benefit from the pirate copy

In my first reading, there is nothing in the *judge's* comments that references anything about "gaining benefit" from a copy, pirated or otherwise. There are going to be futher judgements made about this mater, but this one limits itself to the copying aspect.

Is there actually anything in the US copyright law about benefitting from the copies? There is an "obvious" reaction that you'd only pursue a copier that is actively benefitting - selling the copy cheap down the street market - but the wording I've seen (to date - as I'm asking here for more info) refers just to copies, nothing about what is done with the copies: simply making & keeping copies without permission is the offending act.

Techie went home rather than fix mistake that caused a massive meltdown

Re: Honestly I'm bilingual

Let's just stick to pounds, kilos - and tuns.

Re: Honestly

That depends whether you are controlling the measurements with an entire PC or decided to go with a more cost-effective microcontroller, say an ATTiny.