Re: medical data, identity documents
> The problem is that contents encryption is a massive pain in the neck
More of a pain than random military emails running about in the wild?
5065 publicly visible posts • joined 9 Nov 2021
Not your down voter, but::
> without the need for some sort of key-exchange beforehand, presumably by some medium other than email?
> Try "I'm going to send you an encrypted email, send me your public key please" on the sort of user who is going to accidentally email state secrets (in plain text) to Mali and see how far that gets you
Well, I'm sort of guessing here, *but* before you start sending out state secrets to people, how about a good old fashioned face to face meeting? Then you can exchange keys in person.
Say the top guys met like that and vouched for each other, exchanging keys. Then they go back to their offices and each vouches for their lieutenants, passing on the keys. The lieutenants vouch for the sergeants and so on down the chain. And back up it again, so by the time of the second meeting, to shake hands and drink champers over the agreed contract, the keysets can be passed across again, face to face.
We could give this scheme a name: how about The Web Of Trust? Catchy, isn't?
Outside of the military (who can just order people to swap face to face) and Big Military Contractor (who can order as, really) we can have a nice Wine And Cheese evening and call it a Key Signing Party.
We can have nice little UIs that even top-brass know how to use; heck, even dress it up as a Mission:Impossible style gadget, which shows pictures of all the key owners flashing past when they each each shove their USB stick into the slot. They'll love that.
PS if you just want to send an email to someone, they can just put their public key on the website. To make it trivially easy for Joe Bloggs, just put it it in a clickable link, inside the "contacts" record or otherwise tag it so the browser/email client Just Does The Right Thing.
Thank you for your very complete description of what I said ("email long predates MS") - most impressive ability to list RFCs.
But you must learn to distinguish between someone being responsible for doing something positive (inventing email, inventing encrypted email) and being responsible for something negative (discouraging encrypted email to the point most people don',t seem to know it even exists).
As you have such a comprehensive knowledge of the history of email, you know well that earliest days didn't bother much with a widely-known encryption scheme (we were happy to have email, we trusted most people in the small world before "IT" was invented). Heck, ROT13 was new! (No, it doesn't encrypt worth a damn, but it looked scary to the uninitiated).
You also know that before MS even started including an email client in every install (Outlook proper dated to 1997) privacy had been an issue for while. PGP (1991) was (arguably) the best, providing encryption, signing and the web of trust for key exchange (face to face for best security, of course): wrappers for common mailers were created and the Open Source mail clients started providing options to use it.
Your own favoured clients, pine and elm, allow you to use PGP and/or GPG (the latter coming in 1997).
Encryption was becoming generally available in the Internet email (sadly, proprietary pre-popular-internet systems remained incompatible, sigh).
Then Outlook. Not the best, but increasingly widely used - as time passed, if you weren't compatible with Outlook (and its increasingly large set of headers) then you were in the wrong. More and more non-techies used email and even those you would *really* think cared about signatures at the very least (lawyers and all the people signing contracts) didn't bother to find out what was available. No point in encrypting your emails because chances were the recipient had no idea how to decrypt and Outlook gave no suggestions to help.
Microsoft had the opportunity to follow what the other email clients provided but steadfastly refused.
So, yes, I do hold them responsible: it was a deliberate and knowing action to refuse to support a feature available in the other clients (hey, they didn't own the web of trust!) not even providing their own, incompatible, encryption. That wouldn't be great but at least it would get people thinking about secure email.
You have heard of "embrace, extend, extinguish", well sometimes "ignore, ignore, extinguish" will also work if you have the market share.
PS Apple didn't help either, but again, email isn't proprietary, so - once they started growing in stature again - why would they bother?
Because Microsoft[1] refused to support anything like PGP/GPG (especially in an easy to use, "click here to encrypt, leave unchecked and we will just sign it for you" format) and their bad practice becomes the default.[2]
[1] yes, email long predates MS but it is the practises of mass adoption that are a problem; you can spot the Old Guard and their protégés lurking on tech boards, their public messages properly signed.[3]
[2] See also top-posting, ye gods, top-posting and the apparent inability to insert your responses directly beneath each question in an interrogatory email!
[3] Especially hilarious when you see all the bullshit "if you are not the intended recipient" sigs
"I suspect that sort of traffic will not exactly contain nuclear codes"
Not since Trump sent his infamous GMail bulk CCed "Change the codes to 56789, that one you gave me is too hard to remember" email.
Note the cunning way that code starts *after* the one for his suitcase.
You have any actual *proof* that, say, the kernel is spying on you? How is it doing that? Maybe they modified bash to report back to homebase?
Given that normal ChromeOS users spend their entire time within the browser and extensions, they don't need to fiddle with anything else, do they?
And note that I did not suggest another Chromium based browser, so no claiming they are all using the same core engine and spying that way.
> why do Linux writers take such pride in the Grey-beard scenario
Ah, isn't that paragraph you quoted - and other chunks of the article - Liam sarcastically ripping on that same "if it isn't hard it isn't Linux, what do those youngsters know!" attitude?
Sarcasm? In The Register? Shocked, I tell you.
You do recall correctly.
Although <changes to nasally voice> polite discussions(!) are had whether they rolled up or folded: when closed, the handles are almost as wide as the exposed screen. And they are rigid enough to be easily held in just one hand when opened. Even in a gunfight whilst you are crouched behind some boxes.
Nerdmobile - Away!
> a roll up Kree tablet as seen in Agents of Shield season 5
I see your 2017 telly and raise you the 2000 movie "Red Planet".
They even made it a plot point[1] that these things are (probably) so thin you can view them from either side and spot the view of the mountain range is from the other side!
[1] more realistic than the main plot point - and even that was miles ahead of the key reveal in the same year's "Mission to Mars"!
Always good to see patents accompanied by the crisp diagrams from your CAD designs, created when you built the first test devices.
Ah, misty-eyed memories[1] of the days when patent applications were required to include a physical working copy (or a model, to avoid the office being cluttered with too many 20 yard high beam engines).
[1] hyperbole, ok? Not really that old! Adjusts silk topper, straightens cravat.
If you watched the first video and were wondering why that bloke carried the wooden ramp away after I'd said they could store it in the back of the Cybertruck:
That was one of their first test item "reusable ramps": it had to go back to the workshop for a clean and refurb before being re-certified. Musk promises that, by the end of 2025, they will be able to reach a ramp re-use cadence up to one sidewalk every 10 days.
Meanwhile, standard disposable single use ramps will be sold in packs of nine from your Tesla dealer.
The worst part of this whole affair is that POL would not listen to their own people even as the number of apparent frauds grew to obviously stupid levels.
So drag Fujitsu through the mill but keep talking about the higher-ups in POL who let them get away with that nonsense.
Anyone remember the Horizon about "Future Work" that showed the on-site engineer spending his day, alone, practising archery[1]?
It was supposed to show a life of leisure but looked really bleak. Horizon pre-empting Black Mirror.
[1] this went out before I joined the workforce; at the time I did not realise that archery would be a useful skill for dealing with users.
"Can you pick up the Sales figures, Marvin".
Can I pick up the Sales figures?! Here I am...
... Call that job satisfaction?
(and we know the rest - I'll stop here in case anyone is getting bored of DNA).
> Of those that have used generative LLMs, 30 percent tried it once or twice, 28 percent use it weekly, 9 percent use it once a day and 8 percent use it for work
Adoption levels? For something where the 30 percent could try it for free? Weekly and once a day? And of the rest of them, how many needed to pay for it?
How on earth is that a sane comparison to smart-speakers? Which weren't cheap, most certainly not free, at this point in their life cycle? Heck, if I ever got a smart-speaker, even I would "use it" more than once a day (just telling it to switch on for the news and half-six funny is at least twice a day).
Even the stats are being over-hyped!
To the "More Or Less"-mobile (oops, that's three times a day).
A perfect example of the way this sort of pattern recognition system works (how any neural net works): in ways that make a mortal man scratch his head and day "huh, didn't spot that".
And also a nice illustration of how these weirdities can be a Good Thing: all the people scared of geolocating their house from a random photo can relax.
An image lookup from a StreetView picture (presumably removed from context and with metadata stripped - does StreetView put metadata into saved images?) back into StreetView could well be a useful feature (maybe not to me or you right now, but...).
Training a neural net to recognise a character set is a perfectly good use: I'm sure I've got more than one textbook on Image Processing that has (a smaller version) of that as one of the worked examples[1].
The only bit of your idea that I'd caution over is the procedurally generated bit: you run the very real risk of training the thing to look for odd characteristics that come from the generation code.
The nice[2] thing about ANPR is that you can't lose: so long as the photo is taken of a vehicle on the public highway (hint: choose where to site the camera) any failure to spot a reg number means you've got someone you wag a finger at. Maybe just needs a squirt of water on the plate, or you've caught one of those twats who think using a "fun" font goes well with their personalised plate (or even worse, uses bad spacing and cut-up letter shapes to pretend he got the personalised plate of his dreams).
[1] I'd give a citation but I need something that will read all of the textbooks in my library and then let me ask a question without being able to remember a specific phrase or term used. Oh, and only give accurate answers or "dunno", so cuts out the 'obvious' candidates for that task.
[2] from the p.o.v. of the student using the text book, of course that is all I meant.
> Certainly the label "AI" is overly broad and over hyped.
To quote myself from a couple of days ago:
AI research is a bit of a weird one: one of the old sayings was "if we've figured out how to do it, it isn't an AI question any more"
Years ago, we had no idea how to do this and get it to work, Machine Vision most definitely a topic for AI research.
Now we can see how to do it (well, *one* way to achieve the goal), oh, it is "just pattern matching" so referring to it as anything "AI" is over hyping!
Do not go into AI research if you ever hope to have people quietly applauding your results.
> touring theatre actors are probably not all going to be driving around in flash cars and living the high life
What's with the touring, bring back the Rep![1]
They won't be driving flash cars either, but all the "An Audience With" shows in their later years will allow them to keep on making a living, by regaling us with stories of the landladies they all knew. Especially when you get "Together Again, The Cast Of TV's...": "You remember Bournemouth?" "Oh, no, don't! Her sausages, always like" (whole group laughs).
[1] for the youngsters, the theatres would have companies of players that, instead of touring one show all over the country, one or two nights apiece at each town, would stay put for an entire season. They had a repertory of plays (hence the Rep), a number of standards and then the new ones for that season, and they'd do them all in rotation. Generally rehearsing in the day for an evening performance (with a few matinees of the previous day's piece added, just to keep them exhausted). Standards, of course, meaning a lot of Shakespeare! When you went on holiday, the day was spent on the beach and in the evenings the kids would go to the flicks whilst parents and aunts would "get a bit of culture" with three plays over three nights. The Rep is where we got all our National Treasures from, Dame Judi and the like.
But you tell the youngsters this and they just won't believe you.
That is more to do with needing an extra order of magnitude of components in the skin, with variation of tone and muscle, making the CGI model more complex. Nothing AI about it. Except that calculating such large models needs all the GPUs, just like creating and using the varieties of neural net models does.
Maybe a net model would be a good way[1] to generate a few thousand such complex CGI models, but once that was done just keep on reusing 'em.
[1] or maybe it'd be better to use an improved physics model? But that would take, you know, cleverness compared to just flinging a net at the problem.
> To be honest, I'd do it for $200 as I have no aspiration at all to ever be an actor so to me it's basically a free $200.
And when you - oops, just your likeness - become well-known as "The Face(!) Of Pteparation H" would you regret not getting residuals? Or even free samples (does wonders for tired eyes).
It would be a lot easier to just wait for not famous faces in TV and film to pop their clogs and build the models from their old works.
The ones who haven't set up massive trusts to argue that their likeness is protected until the Disney limitations run out.
But as for you suggestion, why invoke AI (with or without scare quotes)?
We already have good enough rigging for human form models in CGI and can create models to wrap onto those, giving us the monsters, Ultron to Gollum. Even if we still need to have Andy Serkis's students doing motion capture, two hours of "walking on city street" can be reused for background extras. Generating coverings for those rigs doesn't need anything AI or "AI", in fact you'd be better off without: less chance of being sued "because you obviously used my likeness in the training", as opposed to "nah, we just threw dice and fed it into the body generator; you aren't really unique, get over yourself" and use the existence of lookey-likey agencies as proof of that.
Real people in massive crowd scenes have been a rarity since Ben Hur (ah, the glory days of Samson and Delilah and that flood scene...).
The elephants were walking in a ring around the camera, with set dressers throwing different blankets over them.
And would you believe that most of the cityscape backgrounds you've been watching aren't real either?
Good old Uncle Travelling Matte.
> Perhaps he should be grateful something is still consuming the advertising.
So should he pivot over who he sells the advertising to? Otherwise, it won't remain relevant to the readers of his site.
"Do you suffer from blocked IP packets? Is *your* connection timing out? You need to try Nagle's Ointment!"
"You can't surf without Cerf's Solution, The Best A Bot Can Get".
> And it would mean the AI companies would need to create ways that institutions (universities, companies, ...) that want to apply AI to material not available for free on the web would have to be able to add data to their own instance of the model
That mechanism already exists - it is referred to as "pre-training". You just get a copy of the model after it has been fed "generic" material and then keep on using exactly the same method to give it all your domain-specific information.
To illustrate the idea, you can get hold of pre-trained models, of various utility - e.g. https://github.com/onnx/models
AWOOGA AWOOGA AWOOGA
Incoming pedantry alert!
"Smeg" as a brand... not "SMEG as a brand"!
At least, that is what their own website uses in plain-text paragraphs.
I've taken to biting my tongue when people talk about the programming languages LUA, PASCAL, MODULA, JAVA etc but there are limits :-)
But you have to get everyone around the campfire totally newted before they'll all agree to hold an old lens up to their eye. Although you could then try to convince them the next morning that the strange thudding noise and weirdly bright dawn was alien encounter residue.
Hmm, jake's approach seems easier (and cheaper! Have you seen how much Old Red Eye those guys can sink?!)
If you take this copy of the video of that last StarShip flight (it is a very small file so that we can easily copy it around the truth-hunting brotherhood) and zoom from its original size of 640x480 pixels to full HD, you can *clearly* see the alien's *real* version of StarShip phasing in and out of visibility around the so-called "rocket".
After the dummy craft starts to tumble through its own cloud, you can't see the alien craft any more, PROOF that it accelerated away into space by ion thrusters.
> You wouldn't give someone a company car without making sure they were fit to drive so why would you give them a computer without making sure they were able to use it safely and securely?
Car? Start by checking they have a driving licence.
Computer? Strange, IIRC there was much derision around these parts when the idea for the ECDL was announced, yet here we are today...
Or maybe use the Return key for "I've read this" (oh look, it already does! so does the space bar[1]) and if you also want it moved to another folder pop that into a macro?
Ok, order your peon to "pop that into a macro".
Keep the delete key for, um, setting things aside "to be deleted".
Hey, maybe that is an answer? Just (get your sysop) to rename "deleted" as "to be deleted".
[1] on this mail client at least, yours may differ - bit naff if it does!
Dragging the disc to trash was - obviously - how you eject a disc on a Mac.
If your hard drives contents were just being shredded you should have sent it back as faulty and demanded a replacement where the gas-powered platter ejection[1] was working properly.
[1] patent held by Q, c/o HMG: "One swish and all your cares are gone" (tm)
> screen wash & wipers are already invented
As are the scratches from the wipers moving in the same way every time. Nice fine scratches and laser light - oooh, look at the pretty diffraction patterns.
Maybe the emitter window is really funky stuff and way too expensive to make in a sheet large enough to bond into environmental casing and putting it behind a window causes problems of its own (e.g reflections)? Optical issues like those can be solved *but* maybe at too high a cost to make it worth bothering.
So you still have to check whether there is anything that an cause issues with "the obvious solution", hence why I put it in.
The same goes for every other part of putting a doodad into a new use case: can you make it work at a cost we like? Didn't think I would have to spell that out.