Posts by that one in the corner

Sergey Brin promises next generation of Glassholes will be much less conspicuous

I used to love the idea of a wearable display on my specs

It was the dream, back in the 90's, when glasses with displays were being demoed - and sold, from multiple companies.

To be able to wander off and keep an eye on the build status (without pulling out the mobile phone, or even the PDA as we had back then) or be able to look at someone else's PC then bring up the docs from my PC - even better, whilst standing there, edit on PC via the single-handed keyboard (and be able to use my favourite editor, not their monstrosity of a character mangler) and send the change over. All without rudely walking away (which would give ample time to forget what I'd just seen, let both of us be interrupted by other people etc). So on and so forth, nerdvana. And no, even VNCing into the PC from a mobile phone is not the same!

But nowadays, the insistence that the specs are surveillance devices as well and other stuff shoved in ss well, making them "devices" and not just peripherals that you can use as you wish...

Well, first it seems the prices haven't come down (there are cheaper display-only gadgets that turn up lkke clockwork, but they generally seem to forgotten what was known three decades ago about basic optics).

And worst, the general attitude towards using them has become untenable.

Even leaving aside any quite acceptable anger at being spied upon, the general stance of "they look like assholes and we are in the right to treat them as such" pervades everywhere.

Oh well. I'm no longer looking over colleagues' shoulders, trying to figure out why it does different things on my PC but works for him (grrr), so the office nerd dream is put aside. Shame, that would really have been a useful display form factor: I did have the chance to try a few units, including a combat helmet mounted one, complete with netting; daft thing is, you'd probably get less idiotic crap out on the streets wearing that one!

Although I would still like a specs display, to wear out on the street, to help jog the old memory. That was one of the Good Ideas for a specs-mounted display.

Windows reports two CPU speeds because one would be too simple

Re: I thought..

> whereas a BBC B slogged along at a solitary single megahertz

2MHz! Though there was the 1 MHz port when you turned it upside down (and you could also attach a second processor via the Tube, IRC one was a 4 MHz 6502-compatible).

Although perhaps I should have kept quiet, as you were impressed by the 1 MHz versus the Elan and that doesn't seem as impressive now).

Please, don't go there.

Trying to get used to those modern prefixes, working to find a couple dozen spare bytes in tiny AVRs one minute and spec'ing PCs the next, I already have the total hibi Gibi's.

Apartment living to get worse in 5 years as 6 GHz Wi-Fi nears ‘exhaustion’

I can see the appeal: who needs 'phone stands when you can just click in to the fat, orange and inflexible ether cable!

Get the pipe-bending right and you can have a loop come off the wall and hang the 'phone at just the right height to go hands free whilst live messaging during "Come Dancing" (no "Strictly" 'round these parts, me'boy).

Re: Fiber

> One WiFi router for an entire building is nonsense anyway

Who is suggesting that?

From linked article:

>> We included every 6 GHz Wi-Fi access point and active client device (e.g., smartphones, laptops, tablets, TVs and other connected devices)

No mention of one router per building.

Government sponsored faraday cages

Shield all the apartments from each other; run your 'phone calls over the WiFi as well (or send the persistent 'phone user to stand outside, with the smokers).

Linus Tool Tips tried carbon-infested paint, which is good for moody teenagers, but small mesh chicken wire in the plasterboard would be an interesting change to building regs,

(On the downside, slum landlords will be charging extra for the water soaked plasterboard)

Re: And Gamers...

Well, I do recall an episode of The Gadget Show (think it was that - I hope so, otherwise that makes this even worse) where a Twitch gamer proudly broadcast to the nation that he had dual WiFi repeater boxes to keep his large rig, camera and boom microphone connected.

And to prove it, you could see both sets of antennae, one box positioned on the LHS of his desk, the other on the RHS, neatly arranged to the sides of the extra size monitor.

Re: Run wires

You can easily go thinner if you give up gigabit speeds to every single connection: 10/100 Mbps only needs half the cores. And if you aren't cascading switches onto that piece of cable, 100Mbps can be perfectly usable (your UltraHD with extra wibbly bits 90 inch TV set is not mobile and if you can install that then try to complain about the ghastly costs of hiding full size ether cable run to the fibre box then you get no sympathy here).

But don't go flat. Flat cables get so flat and droopy by not bothering with the twisting, which is not a good idea after a few centimetres.

Builder.ai coded itself into a corner – now it's bankrupt

Re: Wasted potential - there was clearly mismanagement but it wasn't all bad

> However, they were genuinely trying to get building blocks to quickly assemble apps and AI code to convert the adaptations to those blocks from English-language requirements to code. The issue was that they didn't also manage to incentivise the contractors to actually use any of the generated code, because being billed by the hour and given no particular need to use that code they preferred to generally throw it out and write things from scratch. This meant that the ability to work out how close they were and refine that conversion was very limited.

Assuming that is a reasonable description of their process, it raises so many questions, which includes:

They were getting the AI to generate an app, the paying contractors to fix the app (presumably to the original requirements) then planning to use the difference between the two as the metric?

Assuming that the contractors were going to hit the requirements bang on each time? Or were they feeding back in the list of the AI's failings plus (different code and list of its failings)? Instead of letting the LLM "figure it out for itself" and only iterating over the match to requirements. See below about what they may actually be training this LLM to do...

> they didn't also manage to incentivise the contractors to actually use any of the generated code

That sounds like one of the things that needed to have been figured out the day *before* engaging the first contractor! And thoroughly tested as a mechanism before making any attempt to include it as part of the LLM's training protocol. Let me guess, the first year they used in-house coders, who were salaried , informed about the purpose of their work (being part of the training protocol) and were all enthusiastic about "getting the LLM to work" as their goal, as opposed to just chucking out yet another tedious app. Then the contractors were hired and they didn't have the same goals...

No BOFH clauses like "the rate we pay is divided by the number of changed lines"? And this incentive going directly to the individual contractor, not via middlemen (i.e. a contracting firm bigger than a one-man band)?

And we're 100% sure it was not because the LLM was generating gibberish and it was actually genuinely easier to start from scratch rather than try to figure out what was actually going on? Because there is a sneaking suspicion that the LLM is effectively being trained to generate code that REQUIRES humans to correct it and it may be optimising its outputs to maximise that!

> This meant that the ability to work out how close they were and refine that conversion was very limited. They had some spectacular cost over-runs on certain projects, again from the contract side.

Running projects for paying clients whilst they still had contractors in the mix, instead of a functioning AI? (If that was even going to be possible...)

> To be charitable, it did look like they generally took the hit when they took on a project for a fixed price and contractors ending up costing too much

Nothing charitable there, they took on the clients without having a working LLM. if they were hoping to use those clients as part of the training, then that is simply the cost of training. Otherwise, it is the cost of not going back to the client and admitting that they had sold them a service that didn't exist. Isn't there a word for that, begins with 'f', rhymes with 'Lord' (as in "have mercy").

Looking back at that lot, maybe the contractors did have this company by the short and curlies after all.

Re: "to leverage AI tool"

Are we not going to run it up the flagpole and see who salutes?

I miss using the flagpole; those were the days.

AI can't replace devs until it understands office politics

Re: Coding isn't dead...

> . This is inevitable because the endgame for coders (and organizations that produce code) is producing code that produces code,

Also known, variably, as: interpreters[1], compilers, compiler compilers, macros, web page design packages, macros etc etc.

All of which seem to have opened up even more actual programming jobs. Without really killing off the old ones[2].

The input languages get more expressive. Although, for some strange reason every time someone claims their language is "closer to English"[3] the more it is left on the wayside by those languages that take a more rigorous approach...

[1] Yes, interpreters "produce code" from your source code, they just execute the newly produced bit and throw it away - the "produced code" can even be so ephemeral that its representation is just the current state of the execution stack.

[2] What? Well, count how many people needed to use assembler in the 1970s and the count the number today - betcha today's number is larger, even though the percentage is way, way down; and the assembler needed is all the really weird tricky stuff, no getting away with just knowing LD, STO and a couple of variants of JMP!

[3] Yes, English. I can't comment on any system that claims to "be closer to Dutch" or "closer to Greek" because those are all just APL to me.

Re: The Great Replacement

> With an AI, you may start from zero every time unless the AI is feeding off some kind of record.

And you have remembered to make sure to pay the subscription to ensure that the copy of the LLM, that you are carefully training with all the info about *your* product, is totally isolated from everyone else's instance. In particular, the instance that Joe Public gets to use for free/low cost.

And keep paying. To prevent that lovely training being slurped up after your project is done, put on the market and taking business. It would be a shame if a competing product suddenly appeared, based upon all your hard work but with a lower subscription than the one you have to charge to claw back the costs of analysis, requirements etc.

Re: This is what I keep saying

> I've had a lot of success recently chatting through existing processes and turning them into usable docs with the help of ChatGPT.

Good for you, glad things are working out.

But from the outside p.o.v. what you are doing sounds like super-ELIZA; obviously, bigger than ELIZA, with more responses - but then, ELIZA is (in current basic coding terms) gloriously trivial. Is something in between that and the costly LLM going to be suitable, and cheaper? Who knows, nobody is looking at that, use the LLM.

Of course, that use of the LLM is what we used to do by talking to people, which had the result that there were then two or more people who understood what is going on inside the processes, along with all the whys and wherefores that don't get into the resulting documentation - at least, not until a reviewer or user asks for them. Whether this sharing of institutional knowledge is worth the cost of not using the LLM is a separate question.

Re: This is what I keep saying

> write C code to draw a procedural stars with a variable for the number of points

There is going to be plenty of code around for the LLM to have read which already solves that sort of "problem" - there is no need at all for the AI to do anything along the lines of:

> The AI figures out how to solve the problem and meet the requirements.

Your star-drawing code is a perfect example of your own previous sentence:

> AI fetches up a solution which has its origins in work done in the past by a human.

In other words, you are using it as more of a search engine than anything else - ditto the:

> It will also add in the stuff to initialise SDL and create a working mini app.

It is just spitting out the pattern it has seen so often (and which it has probably had reinforcement is a "good pattern to repeat" by its users/devs/trainers).

But then you:

> would just extract the bits I need

So it isn't actually the perfect pattern for you, as you still need to massage the results.

> As a user of AI, you should be able to read and understand the output and know that it serves the purpose and how it integrates into your wider project.

Which is precisely the attitude you also need to take with the results of StackOverflow, or random Github gist or any other result coming from Bing/Google/...

So you are using the LLM as a search engine - great for you.

The question still remains: which gives the best bang for the buck, using the LLM or Google? Oh, not just from your (current) perspective, but from everyone's (energy etc requirements, monetary - say, when they start to charge end-users based upon real costs of dev & running).

Finally:

> You are talking about a wider system, and I am talking about solving smaller logical problems within the system.

The goals of AI applied to creating systems used to all about "the wider system", because that is where the biggest issues lie (it is a bugger to have a GPF, it is life-ruining to not know how to calculate profit/loss needed to run a Post Office branch).

What a shame if the best these LLMs can do for you is the piddling bit.

Torvalds' typing taste test touches tactile tragedy

Re: Out of control

Why it is on the home row is totally a throwback to the oldest manual typewriters: it was really a lot easier to engage CAPS LOCK if you were going to type more than one upper case letter in a row! Well, I certainly thought so (copytyping some stage play pages for a friend at Uni, with the character names and stage directions in caps... Oh what a relief to get back to the machine room and the Hazeltines).

Nowadays, a totally pointless key - except on a tablet or phone, where even the shift key is a locking key (it just unlocks after one letter).

As mentioned recently, if you are unfortunate enough to have CAPS LOCK on the home row, remap it to CTRL for a better way of life. You can then remap the strangely low LHS CTRL to something else - or just leave it as another CTRL for easy access to copy/paste when you aren't so much typing proper words but are just mousing and moving things around.

Re: IMaybe not a piano

> but dissimilar in that hopefully there are no chords!)

CTRL-SHIFT-META-K

Ok, not a chord in that you can't (usefully) press *any* combination keys, just the specific modifier keys plus one other.

More like triggering drones than proper chords.

Although, there *are* explicitly chorded keyboards: the MicroWriter (first saw it with the BBC Micro, then on the Agenda (is the correct name?) PDA. Or the Twiddler.

Is there anyone here who mastered the Twiddler?

Re: Tab key on the right please

Is that keyboards not keeping up with what software uses - or software not being written to take into account what keyboard layouts actually look like?

TAB to step into next field in a dialogue (and then in a spreadsheet) - that is definitely the norm from the WIMP GUI creators' rules that is being followed today.

But was it the norm in pre-GUI programs, built with curses or similar? Right now, I can not recall, for sure, but I don't recall using TAB for anything other than "TAB completion", even away from the shell. Then again, I never used VisiCalc or Lotus 123, so: what did they use to step to the next field? Were they TAB crazy - or did they use the numeric keypad's ENTER key?[1]

[1] BTW, would remapping the numeric keypad's ENTER key to TAB help you?

Re: Bluetooth needed

> Those old model Ms with their DIN connectors won't even plug in to a modern computer.

Funny that, there is a full-fat DIN on the end of my current keyboard cable and - clack, clack - yup, it's talking to my PC.

Maybe 2022 is too old to be "modern"? Ooh, look, one size adapter later and it plugs into the box built from new parts in January this year. Was there a sudden change to PC design in March that I missed?

I can even plug it into *this* (inexpensive - cheap, even) adapter and it talks USB HID*.

* No, real old-fashioned serial kbd to real USB, not one of these weird half'n'half keyboard interfaces they had for a while.

> I think any Cherry Blue switched keyboard would quickly become either "wireless" or "very low profile" if used for any length of time....

One Good Thing about using Cherry switches nowadays is that they are very easily swapped. So if the blues turn out not be your thing, you can get a bag of another colour and make the change. There are more than enough people with expensive microphones making YT videos about keyboard sounds - or just get (borrow, or try out in store for the lucky half dozen who actually have a real computer store to visit) a taster board, with one of each.

You can even mix'n'match - leave the ESC key clicky, to warn when you hit it by accident ("where fid that dialogue box go? Oh, yeah, it clicked away"). Set the Windows key at 20 Newtons to remind yourself not to do that. You can even get fakes, so the Copilot key resolutely does nothing but doesn't ruin the aesthetic like prying it out snd leaving hole would.

chutzpah to duke it out with annoyed colleagues.

Which is where the Model M (or Northgate Omnikey Ultra - even larger!) really comes into its own.

One swift smack from a solid metal based case and the average cheap keyboard will pop its keycaps.

Automatic UK-to-US English converter produced amazing mistakes by the vanload

Re: Still happening

Falling under the influence of the Norman-descended fifth column that remains here, keeping alive the notion of the French words as the posh and proper way to speak. They have a strong sense of irony and weaponised the Daily Mail readers' knee-jerk outrage, making them "bitch about...".

Luckily, aside from their linguistic ambitions, these infiltrators' goals are too confused to regain traction: do they want the return of the Angevin court? Or all the lands claimed by the Tudors? Is it to be the restoration of the English lands, making The English Channel once more merely the inconvenience when travelling from Summer to Winter palaces, or the claim that all of the Norman lands are now to be under the Tricolore - Vive La Manche!

Re: Whoops

> It would have been much better

It would BE much better (nowadays)

> ... smoking isn't as socially cool as it once was

>> On my first trip to the US

We have no information on when that trip took place; if it was in the '70s, smoking wasn't so much "cool" as damn near mandatory! Though back then they'd already have lit up, no leaving it for the break, so maybe late '80s?

> Much better ... "10-minute break". It gives better information

>> a chart saying "Fag break"

*Much* better? Clearly he was trying to insert an informal tone, which is an important piece of information in a presentation, tempering the presenter/audience relationship. To be a *much* better replacement requires maintaining that information as well.

Re: Whoops

Are you okay?

Static HTML pages

> running the replacements directly on the body HTML, and causing lots of page repaints, meant we had to build a REST API

Ok, IANAWD but - immediate response to "static HTML" was "tricksy problem, especially with regexes, but at least you only need to translate each static page once, manually patch any edge cases, save the new static pages in a directory en-us/". Put a trigger into version control on the en-gb/ originals (or set up Make...) to re-process when edited.

But - repaints? REST API? Not cause slowdown? Is this a new definition of "static" page?

Then again, this is a "Who, me?"...

[1] I Am Not A Web Dev

No-boom supersonic flights could slide through US skies soon

Look! Up in the sky, it's - Superfactory!

Megafactory, gigafactory, terafactory, superduperfragilistifactory

Plenty of big airplanes have, over the decades, been built, assembled, maintained and stored in, sheds. Big sheds, but still just sheds. Or hangars, if you are posh. Possibly a Vehicle Assembly Building for the very special flying machines. Other bits were just made down at Works. For pity's sake, Titanic and the Great Eastern were just made down t'yard (sorry, can't do the proper accents whilst typing).

Now suddenly we need to have hyperfactories, pantswettinglybigfactories, mydickisthislongfactories.

Good grief.

Supersonic flight without an audible sonic boom should obviously be allowed

Let's just correct that sentence, shall we:

"Supersonic flight without an audible sonic boom AT OR NEAR GROUND LEVEL..." (and let's assume, badly, that most occupied structures are low enough to count as "ground level").

So from what altitude up *are* we intending to fill the sky with sonic booms? From what I understand, the height of the "cutoff altitude" can be pretty much anything you feel like: NASA test flights have played with moving their cutoff up and down, monitored by mics to check how well they succeeded in keeping the noise on the ground down - or the noise up if they put the cutoff a bit too low. The Boom Supersonic flights are intending to real-time calculate where their current cutoff altitude is and tweak the flight parameters to keep it above - well, above some height I've not yet found (anyone?). Maybe it'll be "useful" to drop that cutoff down to fifth-floor level when nipping over all those big, empty spaces where the houses go more sideways than upwards? Just so long as nobody hears anything, it'll fit the requirements.

Which is all fine and dandy for the voters, but you may have noticed that man-made craft are not the only things in the sky. There are these strange things called "birds" (if you believe they are real) and insects up there - surprisingly high up there (though most of the flying mammals have the good sense to hang around down near the rest of their kin).

Filling the sky with commercial levels of supersonic flight (as we never really give a damn what the environmental effects are of our military activities) may not be smashing our windows every hour, but what is it going to do to birds and insects?

Just like to remind you that our food supplies largely rely on the free pollination services of a wide variety of insects, leaving aside suggestions that we should be eating the things. And everyone over a certain age remembers that in the 1970s, not only did we have to get used to a new currency, but we had to wash insects off the front of our cars at least once a week. No, youngsters, we haven't made insect-proof cars, we've just killed them. Killing (or otherwise disrupting them) more of the insects is not going to work in our favour.

Whodunit? 'Unauthorized' change to Grok made it blather on about 'White genocide'

Re: Quick poll: X users

What about using nitter, because news articles* still insist in giving twitter URLs as their source?

* not necessarily on The Register

Dilettante dev wrote rubbish, left no logs, and had no idea why his app wasn't working

Re: Divers log

> "stones" ... Why humans should be weighed differently than other things confused me

As lady acquaintances of mine have commented on this "you try carrying one around in your belly, you'll be using stones as well!".

PS you'd buy a stone of potatoes as well, or flour, to keep the family going. Dog biscuits by the hundredweight, though.

Re: Divers log

> The Groat (4 pence)

Pennies! The Groat is four pennies.

This is one of The Things About Decimalisation that was drummed into us during the conversion: pence for the new stuff and pennies for old money!

Plus the new coins are not abbreviated or portmanteau'd, which gets rid of the last vestige of confusion of that usage: we have "ten pence" in new, whilst only the crude old stuff has "fivepence"/"five pennies" and "thruppence" (and "thruppenny bit" has "penny" so it is clearly old). "Sixpence" is just one word, if someone hadn't heard you'd always say it slowly as "six pennies" (better yet, "tanner", of course, but let us not risk confusing the modern audience). Yes, for odd values as well: "sevenpence ha'penny", no pause after the "seven" (count yourself lucky you got the first 'e' in the "seven", even in the poshest places).

Which meant we could have sensible conversations about price changes without continually shoving "old" and "new" into every single sentence (allowing them to be kept for emphasis and sarcasm, very important to a British conversation).

Re: Divers log

> one could easily water ski behind it ... they can get there at "warp" speed.

If you can waterski behind a ship being warped* then bloomin' 'eck, who've you got on board, Popeye?

* FYI warping a vessel is an old, old practice: tie a line onto a fixed object ashore, or drop a kedge or even a sea anchor (yes, yes, with a bit of rope attached to the anchor) then haul against that to get your vessel moving)

Re: Divers log

> We would write £1 2s 6d

We would write £1 2/6 or "One and six over a Guinea? Here's a quid and a ten bob note, keep the change dear boy".

Re: Divers log

> - there would be a *lot* of people alive today who had to make this very switch,

Cooeee!

Anthropic’s law firm throws Claude under the bus over citation errors in court filing

Anthropic data scientist Olivia Chen meets legal formatter

According to the pdf linked to in TFA, Chen provided a reference to an article in The American Statistician, with all the usual volume, page number etc.

Maybe I'm just wildly out of touch with the whacky hijinks of American statisticians and the academic qualities of a commercial "data scientist", but boring old UK physicists, cosmologists and the like, use standardised formats for all their citations - we still see the unmistakable (as in, neat, well presented) signs of (La)TeX in their handouts (even if some of them use a GUI frontend, WIMPs) with a quick'n'easy citations list at the bottom pulled from the BibTeX list.

Wouldn't one expect an academically-inclined "data scientist" to have their citations similarly organised, so that any formatting needed for presentation in any document was a simple mechanical task, perfect for any distinctly non-LLM-powered PC to do?

Wouldn't one expect any firm that deals with representing scientists of any stripe to have already encountered such academic citations and have the necessary mechanical formatting just one simple wordprocessor template away?

Probably a complete pipedream, hoping that two sides would be as organised as their claimed professions would imply, let alone that they could manage a simple bit of textual impedance matching without the use of Claude.

Re: Replace "AI" with "computer system" and see how bad they look

No need for the database*: they already had the citation & all its details, it just needed formatting.

* Maybe just a little one, say a simple list of BibTeX references, look the paper up in that to get the citation in a standardised format, which can be fed into some kind of formatter, like, oh, I dunno, a (La)TeX macro that spits out whatever is needed to merge it into the rest of the filing.

Re: Latham & Watkins has implemented procedures

"It's a citation, Jim, but not as we know it."

AMD is Ryzen to the SMB occasion with a bundle of baby Epycs

Isn't that what I just said?

Nobody's claiming creating fragmentation is a new idea (go back to IBM changing the position of a rubber band to increase the speed of your tabulator - for a price).

Ditto the "server runs 24/7/52" - pay more and they'll give you the support. (Even though people have managed to use the Ryzen part 24/7/52 quite happily & they are going to be fine for you, that isn't 'supported'.).

Yes - AMD are trying to create market fragmentation!

By deliberately putting in more, arbitrary, limitations. Which, being arbitrary, are kinda baffling.

So if you want the full RAM, in a server, you buy the 'real' EPYC and give AMD the extra money.

Why wouldn't you buy the desktop part instead? Because the sale literature makes it clear that those are not 'server rated' or some such wording, whatever that means (no matter your Ryzen box at home runs 24/7/52 without any problems). Classic FUD.

Unfortunately, my belief in this idea of any company, especially AMD, being brazen enough to try and get more money by cynically manipulating CPU specs has proven unpopular so far, but until there is a better explanation I'm gonna stick with it.

Market fragmentation.

You wouldn't want to trust your server to a "desktop" CPU, would you?

So if you want more RAM in your server, pay for the Big Boy server CPU.

Google DeepMind promises to help you evolve your algos

Improved on Strassen's 1969 result

Now go and read the linked to Wikipedia page - even it points out that:

1 - Strassen's version has been improved a number of times; his importance is not that he claimed to have found *the* optimal solution.

2 - in practice, you don't really bother with such tricks, because the costs from movement of data swamp the gains from removing a couple more multiplications[0].

So, to be *useful* and (even vaguely) worth the costs of using LLMs (instead of, say, improving upon the techniques from the 1970s & '80s that looked for and found new identities[1]) we need AT LEAST a demonstration that the new algorithm takes into account load/store AND decomposition of larger arrays into the 4x4 subunits AND recombination (including elimination of the excess zero rows and columns for a non 4^k x 4^j matrix) into the final result.

Then we start talking about deriving comprehensive test suites for the code that demonstrates this algorithm.

[0] consider if the demonstration code was written "the obvious way" where element A1,1 was always written as matrixA[1][1], A1,2 as... and all the pointer arithmetic involved. Then the work needed to improve on that...

[1] curses, don't recall the name - there was a novel proof for triangle identities and ... nope, gone, and all Google talks about today is DeepMind; funny that.

Re: Correctness

FWIW for a new algorithm, you don't write a test, you write a proof.

And a proof for Strassen's algorithm, or any of the others that followed[1], is well within the capabilities of non-LLM-type programs (all you need is a good old symbolic maths package - but before we got those working that was also one of the items on the AI researchers' list)[2].

If you are setting an LLM - or any other statistical search - then you'd (hopefully) just bolt on a symbolic maths prover on the end, to generate the carrot-or-stick reinforcement feedback.

[1 ]Strassen was not the last to improve it, he is the one that pointed out it was possible, and useful, to do so

[2] If it works, it isn't AI (anymore)

AMD’s first crack at Nvidia hampered by half-baked training software, says TensorWave boss

There are better ways to do the training?

Software can be improved? There will be new ways to do the training?

For Heaven's sake man, don't say these things out loud, you never know who may be listening!

Why, if the Chinese ever find out ...

Chip bans? LOL! Chinese web giant Tencent says it has enough GPUs for future AI model training

The Chinese are devilishly cunning

You have to have a really devious mind to come up with a solution like this: going back to the text books and thinking about the problem - look here, Reggie, I didn't get where I am today by thinking about the problem!

Everyone's deploying AI, but no one's securing it – what could go wrong?

Re: A candle shop's AI chatbot

The baker has been talked into getting a chatbot as well ("keeping up with currant trends" as he says - incessantly!) but the butcher reckons how he'll wait until the AI is smarter than old Bessie.

US tech titans rejoice in $600B Saudi shopping spree

Re: Phrasing

> "Fighter jet" is referring to an engine.

Yes, thank you, that was my point. Not that the OP will agree with us ;-)

Anyway, enough of this for the moment, it is a bright day, there are two clumps of sunspots rotating into view, time for afternoon coffee.

Still better than the fumes IMO - straight from horse isn't so bad, a few times across the field.

Now, muck spreading days - those you can't ignore! Although it hasn't been very noticeable so far this year.

> if you were only referring to the power unit, you would not use the word fighter on its own in the first place.

Sure you would - the sort of power plant you shove into a fighter aircraft, afterburner and all, looks nothing like the big ducted fuel efficient jobs you see hanging off a freight/passenger plane, nor even the diddly things normally seen on a ten-seater corporate jet. Let alone the pulsed engines that people (well, person, well nutter on Youtube) strap onto a push bike.

Put a selection of those engines into a row and ask anyone, they'll go "fighter", "fighter", "lear", "heavy lift". Not "jet fighter jet" or "jet corporate lease, with option to buy, executive conveyance, jet".

Okay, you will have caught them out on the second one, it was a trick question: that came out of a speed record attempt car, which was never armed and therefore was technically was never a fighter (you can tell by the staining from the salt from the dried lake bed race surface).

> You wouldn't refer to an airliner jet, or a car motor

But you would refer to a car motor versus a mower motor or a chainsaw motor* when talking about components.

We are certain the Saudis aren't just trying to corner the market on spare parts? "Getcha jets 'ere, got 'em in all sizes, love! Got yer fighter jet, yer jet for Jumbo Jets, come over 'ere and show us yer gams and oi'll shows yer a leer jet!"

Those things probably wear out fast, what with all that sand around.

* Ok, many would say 'engine' there; for the oldies even the slightly redundant 'motor car engine' or 'motor mower engine' - you remember the push mower, yes, or the big horse drawn job out in the fields, that filled the evening breeze with the aroma of cut grass instead of fumes.

Qatar’s $400M jet for Trump is a gold-plated security nightmare

> it took some 4 years for the original 747 conversions to go into service and its now been 9 years since the kickoff of the B model and they are still some 3 years away.

Over the last 30 years, how many upgrades and rework has been done on top of the 4 years' of initial conversion? In that time span, those two could well be the Aircraft of Theseus (or Trigger's Jumbo for the more classically educated).

Not that it would excuse the tripling of the time on the replacements, not being a simple repeat job. Especially now Young Jim can't crawl into the ducts like he used to, not with his hip.

After more than half a century, the voyage of Kosmos 482 is over

Alien overlords from the depths of the Indian Ocean

Great.

We've got BLUE HADES in the Atlantic, kaiju crawling out of the Pacific, shapeshifters buried under the ice in Antarctica, Borg under the ice in the Artic - and now Venusian outcasts in the Indian Ocean!

I'm sticking to swimming at the local Lido. Hmm, strange looking wet footprints leading from the pool to the changing room, wonder what

Linus Torvalds goes back to a mechanical keyboard after making too many typos

Re: Nostalgia?

Sorry, you are fantasising - a little bit.

The Selectric was released in 1961, so try 1960s through the 1970s and on a bit after that.

But, yes, a well-loved keyboard, although I recall a couple of the typing pool took a while to adjust from their old "hit the steps like a hammer" manuals (in the those days, when a secretary tapped you disapprovingly with a fingertip, you knew the bruise was going to last).