Posts by theOtherJT 1334 publicly visible posts • joined 6 Jun 2013
It's not as good. Specifically its playlist management is nothing like as slick as the Winamp one was. It also doesn't do such a good job managing devices on the end of the dreaded mtp protocol. I'm using VLC too, because on Linux I can't find anything better, but god I do miss Winamp.
...was "smart" in as much as it had an IR sensor in it that would turn it on if it picked up the heat signature of a human sized warm spot moving around near it.
The reason I bought it was that for some stupid reason the flat I lived in at the time had a long hall from the front door to the living room, and a light switch for the hall down at the living room end, but not at the front door end. So if I got home in the dark, I'd come in, close the door, and be plunged into total darkness until I walked the 15 feet or so down the corridor to turn the light on - during which time I would almost inevitably fall over something that someone else had left there since I went out and I wasn't expecting to find. Oftentimes "find" with an unprotected toe since I left my shoes in the porch so as not to track mud up the carpet.
The IR lightbulb solved this problem perfectly because if I walked into the hall, the light came on. The light would stay on as long as I was in the hall, and then about 10 seconds later, it turned itself off. The amount of "Smarts" required for this are damn close to none, but it was never the less extremely useful.
Compare and contrast with the flat upstairs who clearly had the same problem but took a higher tech solution than I, and could regularly be heard through the floor shouting obscenities at their phone when for reasons best known to itself their smart lightbulb had decided that it didn't want to be switched on - or often the sound of many things being dropped as someone tried to manage the front door, bags of shopping, controlling a small child, and getting out a phone to turn the light on with what was a clearly inadequate number of hands for the job - presumably anything less than 4.
This may be a mis-read of the article, but the implication isn't that it ran "ssh user@host apt update && apt upgrade" or something of that kind.
It's that it ran "ssh user@host" whereupon it got a shell and then continued to do things and if that's the case I'm... well... confused and concerned. It's like it is allowed to take the output of the last command and use it as input to start generating new commands on a different machine. The flow required for things like that to happen is really, really out there.
No, really, how? How did that happen?
"It looked around at the system info, decided to upgrade a bunch of stuff"
How did it make this decision? What on earth code path leads to the machine "deciding" to do something you didn't ask it to? Seems to me letting an "AI" run arbitrary bash commands without sanitizing them yourself is beyond dangerous. At this point the thing is potentially self replicating. Can you imagine the amount of damage that could do if it got lose in a corporate network? Now there's basically a virus hopping around from machine to machine duplicating itself, calling itself again, doing arbitrary shit... that would take forever to clean up!
...is a massive problem when dealing with this sort of social engineering.
What's more likely, that the person I'm about to ask to leave the building is a professional thief / pen-tester or that for the 3rd time this month someone new has been hired and no one bothered to tell the IT department about them or issue them any ID?
Oh, you're here to attend a meeting in room D? I don't see any meetings scheduled for room D, that seems suspicious totally normal because no one in the exec team ever bothers to book rooms they just show up and kick anyone else who's using it out.
Should I question that this person crawling around under a desk might be inserting some sort of network sniffing kit, or is it more likely that facilites have unplugged something when they wanted a socket during cleaning last night - like they do every night despite how many times we tell them to stop - and this person is just putting it back in again?
99% of social attacks work not because we're all so gullible, but because we've lost hope of ever making all the stupid shit that happens every day stop.
It's an excellent bit of marketing-speak audience manipulation, but it's still bullshit. It's like "Right-sizing" when what you mean is "Layoffs" - a much friendlier term that's meant to make us feel better about what's actually going on but is really there to obfuscate the truth. "Hallucination" makes it sound like something we should empathize with. It also implies that it's not the system's fault. Hallucinations in humans are caused by illnesses or the interactions of drugs - things from the outside world affecting the poor human's brain causing them to perceive the world in ways other than it actually is.
That is not what is happening here. The correct phrase is "Making things up" which LLMs do by design and isn't at all because some external influence is messing up their otherwise good intentions. "Hallucinations" in LLMs aren't some disturbing external force. They're part of the design. They're a feature.
...to keep old cars running.
My current daily is from 2005. It's nearly 20 years old. OK, it gets 24mpg on a good day which isn't ideal, but other than that I am yet to find anything that a newer car would do for me that it doesn't and there is precisely zero need for it to be connected to the internet to do anything. I get in and out by walking up to it, getting the key out, pressing the button on the key. I'm going to need to have the key on me anyway to start the bloody thing, so is it really such a hardship to have to take it out of my pocket and press a button?
I keep thinking "It's really old, I should replace it" but... why? What on earth would I be getting - other than better fuel economy - that I actually want? Sure, at some point it will develop a terminal case of rust that will cost more to repair than I can justify, but when I compare it to my mother's car which is 15 years newer, what I mostly think is "Thank god mine doesn't do that" every time it beeps at me for something.
"You're reversing!" ...I know. I just put it in reverse.
"You're speeding!" ...I know. I'm overtaking that bus.
"You've left the door open for more than 2 minutes!" ...I know. I'm loading the shopping.
Every bloody thing you do it beeps! I know the door is open. I know I'm in reverse. I know I don't have my seatbelt on. I know the traction control is off. Just shut up with the god-damned beeping!
This is the biggest problem with these sorts of contracts in my experience.
Your company goes to $vendor and say "I would like to buy a product that does X, Y, Z - what can you sell us and how much discount can we get?"
A sales drone from $vendor says "We have a lovely product DWH2.0* that does not only X, Y, z(sic) but also optionally A, B, C and we can sell it to you at an excellent price"
then someone who's paying attention points out that Z and z are actually not the same thing, but it gets glossed over by some managers who got terribly excited about the idea of at least one of A, B, and/or C despite no one who's going to use to tool caring in the slightest about that.
So now the contract gets signed, with $vendor saying "We promise to provide DWH 2.0 which does X, Y, z, and A in 18 months and for a large sum of monies."
Then the thing shows up and the person who pointed out that z and Z aren't the same thing is proven absolutely right and whoever's in charge - which at this point may well not be the same person who signed the contract in the first place - goes "Oh shit, that thing we bought doesn't actually work. Um, $vendor, how much to get Z implemented as well?" At which point the thing that should have been part of the contract from the get-go is now added in at a massive additional cost and a 9 month delivery slippage.
Sometime during that extended, now 30 or more month, period some other idiot, who got too excited during the initial sales pitch, suddenly realizes that B and C aren't included. "Well, No" - says $vendor - "You didn't pay for those and we did say they were optional..."
...which is the point we discover that they're actually not optional any more, because in the intervening months this new idiot has been basing all their decisions around the idea that the new product would have B and C because they didn't read the actual contract but the nice sales rep sure did spend a lot of time talking about B and C in the initial pitch.
So now we need to get a price for B and C - oh, and by the way, since this project is now well over a year behind schedule we actually need D as well, which wasn't any part of the initial pitch but our requirements have changed because getting this implemented has taken so long.
"Sure!" says $vendor "But D is going to require a major change and we can backport it to DWH 2.0 - but it would be much better you bought DWH 3.0 instead which is finally out now, and will only cost an additional 10% on top of your original contact's per-month price!"
"Oh no! goes basically everyone, that's far too much money, how much for the backport?" To which the real answer is "Fuck if we know, the sales drone pitched it without actually asking engineering if it was possible to do a special custom version of DWH that was basically version 2 but with one of the big features of version 3 shoe-horned in there" so hands get waved and everyone is assured that it'll definitely be cheaper to some extent...
...which after another 12 months of messing about by the engineering department at $vendor to try and get this shit working it's discovered that no, it's definitely not going to be cheaper, but we've still paid for the proof of concept work, and to start again with DWH 3.0 now would be even more expensive.
So now here we are, two full years after the initial contract should have already been delivered and we've paid the full cost, plus a bunch of additional un-planned "upgrade" costs, the thing still isn't here, we're looking at at least another year to get a special weird and impossible to properly support version of DWH 2.0 with only some of the benefits of 3.0 and oh look, 4.0 is out now and that's a whole new thing and 2.0 is going to be retired the year after next so your original contract won't be valid anyway and you'll need a special extended support contract for even more money.
By this point all the people responsible for the initial decision have long since fucked off and all the staff who just wanted X, Y and Z to get their jobs done are many years more entrenched in the (almost certainly Excel based) workaround they put in place to deal with the fact that the old DTAH** system didn't handle Y terribly well, and have added another bunch of stupid shit on top of that which means that when we finally get DWH 2.1.043.misery.q4 actually delivered, it doesn't do what they want it to anyway.
All of this shit could have been avoided if everyone involved actually read the exact specification and contractual obligations but since that has never once, in the history of procurement, ever actually fucking happened we're stuck with this sort of nonsense happening over, and over again forever, and asshole sales drones keep on selling things that don't actually exist to idiot managers who keep buying things that don't actually do what they want.
/rage.
* Disaster Waiting to Happen
** Disaster That Already Happened - because that product also went through this exact same bullshit procurement process
I always got the impression that powershell exists out of some sort of institutional stubbornness and refusal to admit when they are wrong.
Like, after a decade of Windows admins moaning about not being able to properly script common administrative tasks someone at Microsoft finally cracked and said: "OK, fine, we'll concede that we do need a shell for administrative scripting, but we're still not going to ship one of the existing ones that you all already know how to use. We're making our own. And it'll be better. And completely different. For reasons. And you'll all have to learn everything all over again. So there."
Smart guys enjoy complex syntax and high level abstractions.
I think you have that backwards. I suspect that what they really like is simple syntax and as few abstractions as possible, because the more abstraction layers there are the harder it is to understand what's being done.
The problem is that low level stuff tends to have really complex syntax out of necessity because you're having to do all the difficult bits yourself; and while high-level stuff has nice simple syntax, it's frustrating because it regularly won't do what you want it to but rather it does what whoever wrote the abstraction thought you needed.
Getting that balance right is really, really hard.
This makes me laugh every time. You can keep trying to protect your precious bit-stream however you like, but ultimately you're going to have to decode it into something my entirely non-digital eyes can see and my equally non-digital ears can hear and as soon as you do all copy protection is rendered totally worthless.
"Fault detected. Security key rotation failed to authenticate with manufacturer inside mandatory timeout period. Please contact your local authorized dealer to book a service appointment. Drivetrain disabled until security key is reset."
You think anyone stupid enough to put something like this into a vehicle in the first place wouldn't also be stupid enough to brick the whole car rather than let you turn it off?
I mean, in my ideal world I'd be driving a 1966 Ford Fairlane GTA which has been through a whole resto-mod process, but that's an amount of money I just don't have.
...and let's be honest, cars from the 60s and 70s are terrible. Sure, they're cool, but they're slow, uncomfortable, unreliable, unsafe - basically the only good things going for them are they're oftentimes very pretty and usually have good visibility - but most 90s and early 2000s cars still have good visibility too, but also have a decent chance to start when the temperature drops below zero and you might even survive getting side-swiped while driving one.
I get the impression we peaked in about 2002 and the further we've gotten away from the e39 BMW 5 series the worse things have gotten.
They don't give a shit what we want. They are expecting to get paid for this, it's that simple. Why only make money from your customers, when you can make more money selling out your customers to advertising companies as well?
The only way we make that stop is by point blank refusing to buy what they're selling. Hit them in the quarterly profits hard enough and watch them back pedal.
...for not buying any car made after about 2010. I will spend any amount required to keep my existing vehicle on the road rather than put up with this nonsense. The only way we get this shit to stop is to make it clear to manufacturers that we will not now, nor will we ever buy anything with this sort of invasive shit in it.
...and he lost me basically in the first paragraph.
GPUs are not in any way endangered. What they are is commoditized.
Using dedicated hardware to do the kind of programming operations that GPUs are so suited to isn't going anywhere. The difference is that instead of that hardware being a separate card you have to buy and install because that sort of thing is only of interest to people who play games - and not everyone does that, so why pay for it if you don't? - to something that just gets bundled on-die with the CPU because everyone has need of at least some of the features that hardware provides and it's cheaper to include it than not: that's the difference.
You still have a GPU. Just like you still have a sound card, and an (S)ATA card. The dedicated hardware that does those jobs still exists. It's just bundled as part of another chip now.
No one sane wants to be that guy.
The problem is that terrible decision making, corporate egos, and lack of planning keep forcing people to be that guy. Sooner or later you get asked something like "Why is there a cron-job running an incomprehensible sed command over the output of this find command once a day? Also: Why the hell is this machine not fully patched?" Ah, well, that's because SCREAMING_HORROR_2.1.1 on here that outputs into this directory nightly, but it does it in the wrong format because one of our other machines got upgraded to version 2.3, so we updated all the SCREAMING_HORROR_CLIENTS to 2.3 but this one server is still on 2.1 and we needed to modify the output because the 2.3 client can't open it any more, but we can do that by...
...and of course at this point anyone sane would go "OK, stop there. Why not just update this one to 2.3 as well?" and then you have to explain that Dave, who's senior VP of something or other has a Mac, and the vendor that we bought SCREAMING_HORROR from discontinued Mac client support in 2.2 so for him to get the monthly reports we have to maintain the original output in the old format which is why this machine can never be upgraded, ever. Oh and that's also why it's running on an OS that's 5 versions out of date despite it being out of support because versions of SCREAMING_HORROR lower than 2.1.7 require a specific version of libfuckthis and libweresoscrewed that aren't shipped any more, and we tried to backport them, but that broke something else, so now there's also this whole separate vlan with just this one machine on to try and keep it at least somewhat safe which is why there's also a whole bastion server doing some really complicated port redirection on the main network and, how did you find out about this machine anyway?
Absolute fucking bullshit like this is happening everywhere, all the time, and it's always because of Dave or someone like him and the fact that they're doing something really annoying and we can't make them stop because of their position in the org chart - and most large organizations have more than one Dave and they're all doing different bullshit and now, as a wise man once said, all your snowflakes are urine and you can't find the cat.
A while back as part of a previous could transition we were told "Just get everything in the cloud. We have a really good deal with $vendor. Just get in in the cloud. It doesn't matter how much it costs. It'll be fine. Get it all in the cloud. We need to shut down the server room by the end of this quarter. In the cloud. Cloud in the get. Get. Cloud! In the it get. Cloud get. The in. Cloud. Cloud! Cloud! Cloud!" or hysterical management bleating that made about that much sense.
We cautioned that this was going to be very expensive but just got more cries of "Cloud! Cloudy clouds! CLOUDDDDDD!" So ok, we ripped everything out, lift'n'shifted it into the cloud as fast and dirty as possible.
Then the first month bill came in. It was ~$300,000.
All of a sudden it really mattered that everyone "right size their deployment" and lots of questions were asked about why we had massive VM's with 48G of ram running simple workloads, and why aren't we using autoscaling groups etc.
Well... because you said "It has to be in the cloud by the end of the year and we don't care what it costs" that's why. You wanted this costed you should have given us more than three months to do it. Of course everyone just selected the biggest machine they could get their hands on for their existing workload, we didn't have time to do any analysis before we started the migration and you'd have thrown a fit if the service broke, so every team went massively overboard on provisioning "just to be safe". If we wanted to rework everything to be capable of autoscaling that would have required a bunch of service redesign, and no one budgeted any time for that, so it wasn't done.
...is this innovation? It doesn't look like it. It looks like we're just throwing masses and masses more raw compute at doing the same dumb thing faster and faster. Innovation would be getting the same results with fewer resources, or better results with the same.
Well, yes and no. See, people are using "AI" all over the place, including to make decisions - to make choices. People have started getting "AI" (and I'm going to keep using those scare quotes) to analyze large quantities of data and then ask it "Hey, how do we maximize our companies productivity based on this data?" and it gives them an answer.
Obviously any real answer is likely be more complex than this, but lets say it boils down to "The best thing to do is fire these N people." and then the business does it! The "AI" might have "decided" to just ruin certain people's lives because it has no ability to care about them. Sure, the idiots in charge might be the ones who implemented the decision, but they can justify it because the mighty "AI" - which has completely ingested all that data - far more than any human could read and comprehend, so it has to be right, surely? - it said it was the best thing to do.
There are plenty of other scenarios. Police using "AI" to pick people out of a crowd. "Sorry sir, you fit the profile of a criminal to the 95th percentile according to our AI, so we're going to have to search you". Banks using "AI" to decide if they should lend people money. The machines might not know that they're making decisions - because they're not actually capable of knowledge - but by surrendering the human responsibility for decision making to such systems they are effectively the decision makers, and they're not capable of understanding that the decisions they make are having real impacts out in the real world - because they have no idea that there's such a thing as a real world, or the idea of moral consequences.
Before we start getting machines to make decisions like this for us we really need to give them a grounding in what it actually means to do such a thing - and as far as I can see there's not actually any way an LLM is capable of that.
...and this is a massive problem. A much bigger problem I think than people are currently talking about.
There's the meme that says "If you're not capable of great violence, you're not peaceful, you're harmless" and that is an important difference. See, we keep pulling things out of AI training data because we don't want it doing things we don't like. We don't want it producing images of child abuse, or describing people using racial slurs, or... well, pick an example really of godawful behavior that AI seems to do semi-regularly.
The problem with trying to ban data that produces that sort of behavior from the training set - as was mentioned in the article re: the MIT dataset that was removed - is that we didn't teach the AI not to be racist or abusive, we taught it that there's no such thing as racism or abuse. ...which isn't true. A really important fact about the world has been hidden from the "AI" and now it's model of the world is actually wildly inaccurate.
We might not want a real AI to grow arrogance, or self importance, or become rude or abusive or any number of "negative" character traits, but if it's not at least capable of developing these traits in principle then it's not a real AI. To look back at that meme: real intelligence, when virtuous, has the capability of great violence, and chooses not to commit harm. When it doesn't have the ability, it's not virtuous. It's not taking part in a massively important part of the human experience.
Why is that bad? Because if AI has no concept of "good" or "bad", "virtue" or "cruelty", there's no reason why it wouldn't occasionally be evil. Not because it has any "want" to be, but because it can't tell the difference. Leaving this vital piece of information out of the training data doesn't generate an AI that's kind. It generates a sociopath - and it's pretty obvious that what we have now, to the pathetic extent that one could call it "behaving" at all - behaves like a sociopath.
It lies. It gaslights. When it apologises for lying, it does so with total insincerity since it will immediately do it again. It learned the words "I'm sorry" in statistical syntactic relationship to other words in sentences but it has no semantic understanding of what they mean. It's not "sorry" because It's not capable of sorrow. It has no feelings. It is, as many others have already pointed out, just a bigass autocomplete.
...moving these is perhaps a little easier, barring details like having to make changes to their networking connections with the outside world. But working like this is perhaps losing some of the promised benefits of moving to the cloud
Which is why so many of us warned that those promised benefits were smoke and mirrors. For every dollar saved in infrastructure costs we've spent two on porting things and redesigning things - and for five years now everyone in management has said that "This is just the transition cost, once we've moved everything then we'll start seeing the benefit" except that the year this "benefit" is supposed to show up keeps slipping, and we've had to start purchasing hardware again to fill in the gaps.
Hardware that should have been replaced two years ago has started to actually age out and fail and the "cloud native" replacements still aren't ready. We keep finding more and more things we have to buy on top of what was already budgeted for because each SaaS product only does about 90% of what the on-prem version did, so now we have to buy another thing as well and then write some glue to stick them together.
If "moving to the cloud" was going to save us money we could have just stuck all the VM's we were running on prem in AWS about 5 years ago and moved on, but instead we're still redesigning and re-implementing all our services at massive expense to chase some benefit that's always another year away...
We've also massively papered over how unstable a lot of things are. "Sure, this service crashes 3 times a day, but it's fine, it's a microservice. It's allowed to do that. We just restart it in the background and the rest of the stack carries on!"
You go digging into the system logs of a bunch of supposedly enterprise grade applications and bits of them are taking daily, if not hourly, dirt naps, only to be recovered by their management layer and all the client ever sees is that the application got really slow for a few seconds.
Now - one might argue that that is innovation - because we can now survive what would have been a full application crash a decade ago, but I disagree. The application should have crashed and then someone should have filed a bug report and whatever made it crash should have been fixed, but now that doesn't happen.
Systems that were designed for high resilience in massively distributed data centre type deployments - a case where it was supposed to be safeguarding against physical failures of network interconnects, or rack power outages, or other situations outside the applications control - are now being used inside what really ought to be monolithic programs as a way to cover up how shitty their code is. This leads to making the thing massively bigger, slower, and harder to debug than it ought to be. That's not innovation.
...to highlight that the situation with two clients for the same back-end was ridiculous and needed to be fixed.
I mean, yes, from an internal management perspective in terms of time wasted developing two clients, yes that's insane. But then they still did that anyway, didn't they. But the fact that it was possible is surely a feature not a bug. Backends should be client agnostic, surely?
Everywhere I have worked that's caught this particular disease did so because the new CEO wants to believe that we're just a few years away from being the next Big Thing (tm).
We need infinite cloud based scaling because in a couple of years when our product has grown by 50,000% it would be impossible to run it locally in our own data centres.
Of course that never actually happens and you just spend a fortune moving perfectly functional infrastructure to the cloud and then re-designing it in the most obtuse and overly complex way so that it can, in theory, scale to a size that it's never actually going to... but that seems to be delusion.
It's like that cartoon about building a swing. What the business actually needs is a tire on a rope, but what the CEO wants is about 2/3rds of Alton Towers because it makes them feel important.
520k immediately reminded my of my first* computer, a venerable Atari 520ST (which I know didn't actually have 520k of memory, I guess they just thought it sounded better than 512)
I wonder how the two would compare in capability - the ST obviously being a much more complete general purpose computer, but... 150mhz is a *lot* faster, and we're not having to load everything off 3.5" floppy disks.
* Technically I used a Vic20 before hand but it was my dads, and the obligatory BBC-B at school, but the Atari was actually mine.
Don't get me wrong, I think the tax rules for VED are insane. They have done such weird things to the used car market precisely because they're not backdated. Two absolutely identical vehicles off the same production line can have 300 quid of tax difference bet6them because one was registered in 05 and the other In 06. It's mental that tis is what they chose to do.
I get that "nicer" is subjective but I cannot imagine a mindset that thinks that. EVs are horrible to drive. Knumb, detached, heavy... Maybe something like a Porsche taycan or an Audi etron (neither of which I've had the good fortune to try) but as a driver I'd take my 25 year old BMW over any I have tried.
...sort of? Some of them anyway.
I got to borrow a mates polestar the other day - the dual motor with all the toys one - so it has no excuse to feel slow, and for sure it accelerates hard but that's about it. It isn't actually fast because it's got a hard-limiter on the top speed to stop you from cooking the batteries. It also doesn't feel good to drive fast even should you get the opportunity because it's so heavy and tall compared to a normal car (And I hate to think what all the really big electric crossover things are like - the polestar is relatively mild compared to most!) that you show it a set of corners and it does not like it. You can feel that the chassis, brakes, and steering rack would really rather you didn't treat me like this thankyouverymuch.
Eh... yes and no.
Massive hybrid SUV's (which does seem to be most things these days, to be sure) are a nightmare in every case. If you do a lot of urban driving tho, small hybrid cars like the Yaris make a certain amount of sense. They' really just leveraging that electric assist to keep the engine spinning in that economy RPM sweet spot so it's not being stressed doing constant idle->power band changes. The second you take them on a long drive the benefit disappears, but not everyone actually does that. Hybrids can work to improve fuel economy, but most of them don't because they weren't designed to. They were designed to sell into specific tax bands.
That's not how car tax works. I think you are remembering that incorrectly.
If the car is X reg or lower it's still on the old 1549cc and below or 1550cc and above two tier system. It changed to the CO2 based system in march 2001, and then they changed the bands on that system in march 2006 before changing it entirely in 2017 so that you payed relative to CO2 emissions for the first year, and then a static cost in the second year plus a "Big expensive car" add on that you have to pay for the first 5 years.
The problem is it's not the "driving part" that goes wrong, is it? Every single car I've owned has developed some sort of electrical fault at some point. The difference is that older cars you just go "Well, I guess the steering wheel adjustment is staying where it is now then." or "Do I really carethat the climate control in the rear seats isn't working independently?" hell, if the car is old enough it's "I guess ABS was an optional extra on this model anyway so it not working on this one is no worse than me buying one without it..."
In modern cars (and, yes, this does apply to modern petrol cars as much as EV's) they have an annoying tendency to go "Satnav's not working. I shall sit here and refuse to co-operate until you go through a 15 step dance with the infotainment system to put it into Maintenance mode so you can drive it at a maximum of 30mph to the nearest manufacturer approved service centre"
It's no good the batteries and motors working perfectly if the bloody computer system that runs it all has sat down and had a tantrum because the sensor that tells it the boot lid isn't properly closed has decided to pack up.
But it can't, that's the entire point. The OS has found itself in a situation where we're off the edge of the map. Something has happened that's totally out of bounds and we're now in a situation where doing anything will potentially lead to totally undefined behavior. We can't offer the opportunity to do anything because doing anything could cause we-know-not-what to happen.
A BSOD Isn't just a crash. It's the crash. It's an admission that the OS has totally lost control of the situation and all it can safely do is halt and call for help from a human.
While OpenTofu is packed full of laudable features, its evolution is difficult to see as anything other than a march away from Terraform
The further and faster we march away from Terraform the better. I have never come across anything so fundamentally mis-designed in my entire life. The idea that a supposedly declarative language just falls over it's own shoes and starts crying the second that the end-point it's targeted at doesn't look the way it was expecting it to look is utterly unforgivable.
I do. I want that. A lot.
Look, I love driving. I'm a firm believer that rear-wheel-drive is correct-wheel-drive, and that the correct number of cylinders for an engine must always be at least 6. I take my motorcycle out for weekends just for the fun of it. Back when I could justify having only 2 seats I drove a beautiful little roadster and I would take that out just for the fun of it too.
...But I also have to occasionally drive between Oxford and London and I do not want to because both of those places are absolutely fucking hateful places to try and drive in. Unfortunately between where I live and where I need to be the public transport sucks - doubly so given what they've now been doing to Oxford train station for what feels like the last 200 years.
I really, really want - on those occasions that driving is about to make me utterly fucking miserable because I'm going to be doing it in stop-start traffic at an average of 4mph - to be able to press a button on my dashboard and go "Ok, car, I'm going here, you take me and I'll get one with some light reading, ok?"
The self-driving tech might be 99.9% reliable, but when the 0.1% can kill people, it's not enough.This is being done completely the wrong way around. The driver should be driving the car, and the technology should watch them
I get what you're saying, but in a lot of ways the same problem applies. What happens when the driver is doing the right thing and the technology decides, incorrectly, that they're not? A lot of driving decisions have to be made quite extraordinarily quickly. We're all so used to it it doesn't seem weird, but when someone is joining a fast flowing road you're already on what do you do? Speed up and get past them? Slow down and let them in? Change lane to make space for them?
You make those choices based on experience and normally have to do so in the space of a couple of seconds or the opportunity to make some of those choices goes away. Didn't decide to speed up fast enough? Can't now, there isn't time. Didn't decide to slow down fast enough? Now the guy behind you is too close and it's not safe. Didn't decide to move over fast enough? Well, now there's someone too close in that lane and you can't go there anymore without endangering them.
One of the last things I want when making a decision like that is the car second-guessing me. If I start to try and pull over and it goes all "Lane departure warning noise!" and tries to straighten the wheel back up because it wasn't expecting me to do it, that is not helpful because in the time it wasted that choice may now no longer be available to me.
When systems like this don't work 5-nine's of the time they're worse than not having them at all.
Absolutely this. The thing that my motorcycle instructor was always keen to drill into people was "Other drivers are looking for cars to a lot of them you might as well be invisible because they're not expecting to see you. Do everything you possibly can to remain visible"
It's much easier to see a flashing light in your peripheral vision than a relatively small moving object that you weren't actively looking for in the first place.
I'm not sure systemd represents a more modest change.
The way I imagine this working in a fairly broad sense is
* A-root is mounted at /
* Snapshot / as B-root
* mount B-root on /B-root (and the various bind mounts)
* chroot to /B-root
* apt-get update && apt-get upgrade
* update fstab to mount B-root as / at next boot
Now all you need is a little script in initramfs that if the machine fails to boot say... twice in a row... you update fstab to mount A-root at / and now you're back where you were before the update.
If you wanted to go further you could define a set number of snapshots of root the same way something like rsnapshot does and rotates them out so that next time we snap C-root, shuffle them down the alphabet and repeat so there's always N good versions to revert to plus one experimental one.
That's all pretty well understood unixy stuff - hell you want to go old and slow with it just use two different physical disks and rsync rather than fancy snapshot capable filesystems - it'd make the update process take quite a long time, but... eh.
For sure, you have to reboot the system to get your updates applied but I'll take that for a desktop that gets rebooted basically every night when I turn it off since machines take all of 10 seconds to boot these days.
Systemd changes a lot more about how you go about doing every day tasks, which I think is the real reason so many people hate it. It's different in ways that don't seem to be obviously helpful just different for the sake of it. Whereas this sort of A/B updating could literally be a wrapper around apt and from a user perspective they wouldn't even notice the difference except that their disk usage goes up a bit, and disks are mostly pretty big these days.
Edit: This example does of course presuppose that you don't have a terribly complicated disk layout where things in other partitions need to roll their own snapshots, but that's still something that could be accounted for - I just didn't want to make this post even more of an essay.
But that's precisely the point, isn't it? It used to be a job that basically anyone could get because they were looking to earn a few bucks in the school holiday, or save up for a year after graduating so they can go bum around the states for a bit before going to uni / getting a proper job... and of course it still is that but those kids are now in competition with people a decade, two decades - more even - older than they are for that same job because they used to be factory line workers and they've been laid off. What job is a 50 year old dude who's done nothing since highschool but work a factory line going to get? He's got to work, he probably has a mortgage, maybe kids to feed - this guy's job has been wiped out and he's not qualified to do much else. So now he's in competition with potentially his own freaking kids for that burger flipping or pizza delivery gig - and this suits the pizza delivery and burger flipping industry just fine, because he's so fucking desperate that unlike the kids who did those jobs 30 years ago he'll put up with a lot more shit without quitting. They didn't need the money to anything like the extent that he does.
and yet, most people who want to do still work.
Yes. And they get paid relatively less for it than they did in the past when adjusted for inflation - doubly so when you do an adjustment for real inflation like food and rent rather than the cherry picked kind that economists like to do because it makes the numbers look less bad.
We're not "out of jobs"... yet - that's still a thing that can happen, but it's probably quite a way off - but by putting so many low skilled and unskilled workers out of work we've massively increased competition for those remaining low and unskilled jobs. Life at the bottom of the job market is brutal now. It isn't possible to live on the wage of someone who works stacking shelves or waiting tables any more. It's not just about "having a job" it's about that job paying enough to live, and providing the sort of security needed to do other things, like raise a family. The absolute decimation of the bottom end of the job market without any compensation in the form of state benefits is driving a lot of people to the wall.
It's already happening. Hell, it's already happened to a lot of people.
50 years ago you could drop out of highschool and get a job working a factory line by showing up and asking for one. My dad did. Both my uncles did. That simply isn't possible today. Those jobs already went. The supermarket checkout kid - another job that basically anyone can do - is disappearing at a rare of knots because self-checkout systems are so much cheaper for the companies. Go to my nearest out of town sainsburys and 25 years ago they had a line of 30 checkouts, all of which manned by someone. Now there are 4, and a ton of "self service" aisles that are constantly startled that you put an item in the bagging area - Oh, and those checkouts that are still there are mostly staffed by pensioners who have been forced back into subsistence type work for a variety of reasons so the kids leaving school without any qualifications can't get those jobs either.
The switchboard operators are gone. The typing pool is gone. Hell, the janitors are gone - we replaced them with cheap labour from outsourcing agencies - so instead of every business having a couple of employees who get full benefits, there are a couple of dozen local agencies supplying contract cleaners who get no benefits and have to work more anti-social hours because they're doing shifts across multiple businesses.
This isn't a new problem, most of those jobs have been gone for decades, but it's one that we refuse as a society to properly acknowledge. Everyone says "Technology puts some people out of work, sure, but it also generates new jobs so that's OK!" And it is OK, for someone like me. I have both undergraduate and postgraduate degrees. I got one of those new high-technology jobs. But the highschool dropout who could previously get a job working the line, or cleaning floors, or serving customers, or any other number of low-skill jobs that used to exist now can't They're not going to become a systems reliability engineer, or a network architect, or a full-stack programmer. A lot of those people didn't drop out because they were lazy, they dropped out because like my father they had severe learning difficulties and they simply weren't ever going to graduate. The jobs technology created did not replace the jobs they lost. They're just not the same category of employment.
Even if they were lazy, that's hardly the point, because the destruction of jobs keeps coming faster and faster and everyone gets squeezed harder. Jobs that you could get on a couple of GCSEs now won't hire unless you have A levels. Jobs that you could have gotten with A levels you now won't get unless you have a graduate degree. The concept of "hire and train" is being absolutely ravaged as more and more people who could have decided "You know, I'm happy working this shop counter because what I really want to do is spend time with my children and the hours are good here" are getting forced out of those sorts of jobs. The competition for even entry-level skilled work is getting fiercer by the day.
Not everyone can be a doctor, or a lawyer, or a - here's one of those jobs that was created in the last 50 years - Software Platform Engineer. There's plenty of people I went to school with as a kid who simply are not capable of doing work that complex and it's ridiculous to think that they're going to be trained to. Once we've destroyed all the jobs people like that could once have gotten what do we do with those people?
The current plan seems to be "Berate them for being lazy and/or stupid and constantly try and take what little dignity they have away from them." This is a very bad idea. Quite apart from the fact that amongst those academically "failed" kids there might be any number of great artists, or musicians, or athletes - but we'll never find out because they have no opportunity to pursue those avenues - we're breeding an entire class of potentially multi-generationally unemployed people who are getting more and more angry and resentful - which, frankly, you can hardly blame them for.
I'm not a luddite - I actually quite like my high-paid tech job. I like the fact that there's been innovation that lets me do that kind of work. I'm certainly not saying "Smash up the steam-engines, tear down the factories" or any modern equivalent there-of, but we need to provide for the people in our society that this technology is putting out of work, and right now it does not look like we are doing that.
Obviously UBI needs to be funded from somewhere - and you are absolutely correct that the status quo is "Money flow up, money stay up." No one who benefited from the massive productivity gains we got out of mechanical automation, digital automation, globalization - and I'm sure that AI automation will be the same - is in any way inclined to give that money back. All we can hope from the likes of Mr. Altman here is that they actually believe enough in the message they appear to be pushing to use their considerable wealth and status to force the government to take the money back... and then use it to provide a basic income for all the un and underemployed people we'll be left with.
I'm sure that there are a great many truly excellent programmers in India, they're just not working for the $9 an hour outsourcing outfits.
We've got several Indian programmers, some of whom even work from India, at my place of work and as far as I know they're all very good at their jobs. The kicker is that they work for us not some outsourcing agency we hired in to do the job on the cheap.
