I ain't paying for Linux.

You can really, really, really stop trying.

Especially when all I'm paying for is someone to recompile a new package on an old machine that was working absolutely fine before they decided to "cut off" package updates.

This is one of the reasons why my preferred package manager is ".\configure; make; make install".

I installed Ubuntu on three machines last week - and none of them are commercial machines, none of them need a subscription or account, and all of them had the packages, MOTD, etc. associated with these services ripped out.

By all means offer a corporate support service. But if you want to do that, don't spam your free product with adverts for your paid-for product. RedHat/CentOS/Fedora all over again.

I would literally rather roll-my-own if this is the way other distros go too... because then maybe I can turn off most of this systemd shite too. Spent 20 minutes the other day working out why I couldn't just kill GPSd and move it to another port after having changed the configuration - and it's because systemd ingratiates itself even into simple services like that to auto-start them as you touch the port you intend to use or insert a device, and that uses entirely different configuration files. Did we learn NOTHING from Autoplay?

All Ubuntu pushing this stuff down my throat constantly does is put me off using Ubuntu entirely - including professionally.

It is ironic that I started many years ago with Slackware as a desktop, then moved to Slackware for servers (predictability and control), then changed to Ubuntu for desktop (simplicity), then changed to Ubuntu for servers (simplicity and control), and now would really move back to Slackware for servers (predictability and control) and other things entirely for desktop.

My entire reason for using Linux is thus:

- I want things to just do what I say, and work when I do that.

That's it. That's the one, sole reason for my preference in that regard. If I kill a process, I want it to die. If I change a config, I want it to take effect immediately without having to reboot. If I want to rename a device, I just want it to happen (I once spent a day trying to rename a joystick device to always present as the first js0 device under systemd, and at the time it simply wasn't possible, it always just did whatever it felt). If I want systemd or an alternative, I expect to just be able to choose. If I want to install my own local DNS server, that's what happens.

Don't even get me started on those software that demand I install them via snap/docker/etc. as their only way of doing so.

Play ball, Ubuntu, or lose your customer base. Not "signing up" to get an extra two years and spamming you incessantly if you don't.

Because you think that the landline system was somehow massively distributed and uninterruptible, but that cellular, satellite, fibre, radio, etc. aren't?

If anything it would be even harder in the modern age to take out communications - previous generations had the landline and public radio (transmitting from large, obvious, powerful towers) and that was it.

Now you can use cellular, satellite systems (Starlink), form wifi mesh networks, use Pringles-cantenna long range wifi networks, use underground fibre, microwave interconnects, copper landlines, radio-modems, etc. not counting what the community itself can come up with locally to connect that one local site who does have a connection to everyone else in the town.

Landlines are inherently reliant on power, for a start, across dozens or hundreds of paths in the connection. They have also been almost entirely VoIP on the back-end for decades now. So much so that my landline provider is telling me that they are removing the voice capability from the line and it'll just be DSL, then they'll move my phone number to a SIP connection (and I don't use it, so I don't care). It's called "Digital Voice" in the UK and it's going to happen to everyone in the next 5-10 years.

And though many of the above methods are interruptible in a war (e.g. satellites), you have to take out SO MUCH MORE hardware than just a few major exchanges.

The BT Tower in London used to be part of a network of microwave towers for exactly this reason. It was obsolete by the 60's because of satellite and fibre networking. It's been basically abandoned for decades. There is a similar tower near where I live, that formed part of a national point-to-point microwave network in case of an attack. They are all defunct for that purpose and have been for decades. They're just radio masts and cellular towers now.

Because the Internet did one thing - it put everyone, no matter how they connect, onto the same single network. Fibre, wireless, copper, radio-modem, satellite... you're all talking IP and literally talking using IP (VoIP, etc.).

And it no longer matters how you join that network, so long as you have ANY connection whatsoever to another node willing to share.

In the case of a war, landlines are going to be the last target mainly because it's just not worth the effort. Until people start bringing down satellites, you're fine. And even then - at that point connecting to the Internet will be the last of your worries, but it will still be far easier than it ever was.

Just because "other companies" do it doesn't mean it's the right thing to do.

Just because "it can be done that way" doesn't mean it's the best way to do it.

Just because "Excel can do the job" doesn't mean you should be using Excel, that your first thought should be Excel or that another tool wouldn't be more suitable.

Just because "The team's been doing it that way for ages" doesn't mean that someone - for instance a guy coming from another successful Formula 1 team who don't do it that way - can't improve things and actually stop nonsense like that where unsuitable tools are used "because that's how it's always been"

As the articles on this point out - while you can dump information into Excel, none of it is live and up-to-date and it doesn't contain anywhere near the information you might need to run a successful multi-million-dollar Formula 1 team efficiently.

These guys are engineering down to microseconds and nanometers, they have millions of dollars sunk into complete efficiency of every part, every design, every usage, every fitting, every staff action, through a full supply chain, including unique, one-off, cutting-edge designs, products, software, analysis, etc.

And some prat is still carrying around an Excel spreadsheet which is highly suggestive that nobody has ever BOTHERED to find a better way (because there will be one!). Literally one line of quote tells you why this is: "Impossible to navigate and impossible to update". The guy came in, found it unsuitable for the purpose and questioned it. It also had real-world, knock-on effects on their actual Formula 1 workshop and performance: "Williams missed early pre-season testing in 2019. Workers sometimes had to physically search the team's factory for parts. The wrong parts got priority, other parts came late, and some piled up." It's right there in the summary.

And all you get is a bunch of office workers and their IT departments on here going "Oh, well, Excel's fine, it's not even a million line spreadsheet being abused as a primitive multi-user database, pfft!".

I wouldn't tolerate a 20,000 line Excel "database" of products, parts, inventory, assets, etc. and I don't work anywhere near Formula 1 - I work in an office. I'd be ripping that junk out and replacing it with a proper system, and have done a thousand times for a thousand different things in my career.

At the level this guy is working at, with the experience from OTHER F1 TEAMS he was in charge of, and that he's in charge of this one and particularly its operational side... he's perfectly entitled to be taken seriously and it's perfectly laughable that such a team was still using Excel and think it a suitable tool with demonstrable, actual, real world impact on them doing so.

You could read the articles about it and discover that it leads to delays and all kinds of problems with parts supply, and that he's now in charge of exactly the team that does such.

Seriously, do you need to even ask?

Where's the profit in keeping people alive a few years more when they can get asbestosis and spend a fortune on healthcare?

I've said before:

Anyone who got to that point knows that the original wallets will be watched like a hawk - by users and governments alike.

But at any point they could have just joined in a bunch of smaller miners and made an absolute killing with a properly anonymous wallet that isn't being monitored, live a life of luxury, and never have to touch the original coins at all.

Also, if the original creator was, say, funded by a government agency somewhere - even just for research - they wouldn't be allowed to touch that original money, but they would be able to discretely mine their own elsewhere once they realised it was going to take off.

Chances are that the original creator(s) is dead, a secret millionaire, and has no need of or access to that original wallet.

Never, ever, ever, ever, ever, ...

Use software in breach of its licensing agreements.

For instance, it's perfectly fine to use some software a self-contained DLL / library that you publish the full source code to in compliance with its licence. Then your proprietary code can call that DLL as much as it likes so long as it uses the registered interfaces.

LOTS of companies / products do precisely this - you have one in your house right now, I guarantee it. They publish the GPL and the list of software used in the back of manuals for EVERYTHING nowadays.

But that's not different to using, say, the PhysX libraries, Unity engine, DirectX or anything else... whatever you use you need to track licensing, draw boundaries and make sure you're compliant.

One of the major omissions of source-code control systems like SVN, git, etc. is that they DON'T TRACK LICENSING. And they should.

How about this...

We make better quality screens, cheaper.

I'd far rather that than literally ANY other gimmick like folding screens, see-through screens, etc.

If you can't make it cheaper or better (including more resilient, brighter, etc.)... then don't worry about the screen. Find something else to do instead.

Gimmicks don't sell me on ANYTHING at all. I'd far rather have a basic, non-gimmicky thing that was slightly better and/or cheaper.

He incited people specifically to obtain unauthorised access to classified military systems and publish it to himself, where he then disseminated it to the world, without redaction, analysis or reasonable protections.

He probably got people killed by doing so.

Sure, there were things in the leaks that were worth whistleblowing, but an awful lot that should never have been made public (which journalists are good at dealing with and separating so they... don't end up being locked up!).

Whistleblowing isn't the same as "just publish everything you can get your hands on, no matter the secrecy or relevance of the documents". If it was, secrecy would be pointless because everyone would claim journalistic / whistleblowing protections, even enemy states.

Journalism, no matter what impression the tabloid press might give it, comes with responsibilities and most professional journalists are extremely responsible. Assange was not.

We are certainly missing something. Something that doesn't just follow on from rigorous logical operations.

We are not a Turing machine. And Turing told us that any machine we have is equivalent to a Turing machine in its capabilities.

We are capable of doing things that no Turing machine can. Intelligence resides in that missing piece of the puzzle - where we can do things that even the most powerful computers cannot. And it just so happens that we know there are things Turing-complete computers cannot do, and we're able to do some of them. That's not a coincidence.

We don't need to get mystical and spiritual (and the whole "consciousness influencing physical reality" is far too unscientific for me). There are things we can do, that computers cannot do - even in theory. Hence current AI cannot do them no matter how complex or fast we make them.

There was a hope that some quantum mechanism might be at work - we have observed quantum-level interactions in neural cells, and quantum physics acts in such strange ways that maybe a quantum machine might be "more than Turing-complete", like we are. However, I believe that our current generation of quantum computers, for instance, are still considered "only Turing-complete" and nothing further.

It's all very hard to define and lock down, but we're missing something critical that no amount of scaling-up is going to solve.

It's not magic, or inspiration, or courage, or consciousness, or a spark. There is a physical, technical thing we are missing that neurons and the brain etc. are able to do that a mere Turing-complete model cannot do, ever.

And even that might not be the complete picture. But it's definitely something different, definitely something we are closer to defining, definitely something that computers are missing, and definitely something that could be a major factor in our failure to do more than statistical automation.

Why should anyone waste their time when the software you're using is inherently out of date, obsolete, unsupported and quite clearly dangerous to use.

There's no more point filing CVEs against ancient PHP than there is filing CVEs against Window 3.1 or 95. We know it's broken, beyond repair, will never be fixed, and should never be in production use. Why bother to catalogue HOW broken it is against pretty much every attack you want to try?

7.4 has been "security fixes only" for over 2 years and end of life for nearly 18 months. Publishing CVEs for it only encourages people to keep using it, but make sure they patch those CVEs, which are likely to be nothing more than the icing on the cake in terms of actual obvious holes in the software.

Developer time, testing time, etc. is limited and wasting it on literally obsolete software that is officially stated as "A release that is no longer supported. Users of this release should upgrade as soon as possible, as they may be exposed to unpatched security vulnerabilities." doesn't help anyone.

"Personally, I find taking calls through Bluetooth to be more distracting than using the handset"

And studies have conclusively shown multiple times that having a conversation with someone who's not in the car (and thus doesn't pause to let you consider responses, etc.) is far more distracting than radios, passenger conversations or a variety of other activities.

It's literally nothing to do with call quality or speaking. It's to do with the person you're speaking to being remote to you and it severely affects your concentration and/or ability to pass basic tests while it's happening.

The real question, then, is why CB radios were ever tolerated in moving vehicles.

And, indeed, there are levels of window tint that are not permitted, including anything on the front windscreen. Those rules are throughout the UK and EU as far as I know, and police even have "tint meters" they can use to determine at the roadside if you're over the limit.

Those rules, however, long predate speed cameras or mobile phones.

Current "AI" isn't AI. It's just statistical automation.

It cannot infer, which is why it suffers from "hallucinations" which I think I more accurately call "superstitions". It's fed a ton of data, it makes a spurious correlation between that data and the expected results, it deals with new data according to that spurious connection. No different to people thinking that wearing their lucky socks made their team win, so they all wear their lucky socks every game. It takes a LOT of untraining (basically until all the relevant previous training is then in the minority) to get it to stop doing that. And then it can only really act on another superstition that it comes up with.

Statistical automation can be amazing, it can really help the world. But it's not intelligence. And - again - if your job can be replaced satisfactorily and safely by automation, maybe it should. Why should you be slaving doing something that a machine can do far better, for far longer, with no downsides? The problems come when it replaces a job where that's not true (e.g. self-driving!).

Every job lost through automation is a movement towards universal basic income too (of which there has never been a unsuccessful trial). The fact is that many of the jobs that automation replaced have gone forever and nobody really misses them. People doing dangerous, heavy works now replaced with tools and vehicles. People just moving a box from one conveyor to another. And so on.

But around it are still jobs that can't be replaced - from the more intellectual ones to the ones that don't require a ton of thought, just basic human senses and flexibility.

If your job can be automated, you need a new job or to get better at it. Because you can also be replaced with ANY OTHER HUMAN willing to do that job for a penny cheaper. And there will be cheaper, younger, less experienced humans who will do it just as well and want to do it.

AGI doesn't exist, and there's nothing even suggestive of AGI in any of the recent "advances". It's all just the same 1960's schtick that we're given - if only we had more money ("trillions of dollars"), more power, more connectivity, more training data, more time, etc. etc. then intelligence will somehow magically appear out of the data spontaneously. It's never been true, and isn't true now, and probably isn't a truth at all.

https://support.google.com/chr...

ChromeOS devices receive 10 years of updates.**

** For devices prior to 2021 that will receive extended updates, some features and services might not be supported.

Sorry, but if you buy "unofficial" Chromebooks, then yes. But I've watched the support life of official devices (not just Google, but HP, Lenovo, Acer, etc.) increase from 5 to 7 to 10 years in the last decade or so. I ran a school entirely on them (even staff). That wasn't through choice, it's because they couldn't afford anything else, and it damn well worked.

Most Chromebooks, especially those you would buy for education or enterprise, have a more than adequate support life and your users will reject them for speed/hardware (e.g. old chargers) far quicker than ChromeOS will stop working on them.

Yep.

I have PoE cameras, with inaccessible cabling, wifi backups, notifications for obscuring or losing contact with the cameras, on-board SD, local and remote cloud backups of footage, plus detection alerts to my phone (because literally the only person who cares about my property is ME, so the only person worth informing is ME, and I can then check the cameras and decide if escalation is necessary or it's just the cat).

Have I spent a fortune? No. Just cheap "smart" cameras. IP68, fitted and cabled by myself and I also have PTZ models, 2-way calling to them, external lights and alarms on them if I press the right button in an app or configure their "smart" junk, RTSP streaming/recording, etc. The most complex part is the internals to power them via PoE (which is also backed up not only by UPS but by an ATS to switch to solar power banks).

I spent almost nothing on the cameras themselves.

I fit CCTV as part of my job, working IT in schools - mostly CCTV is useless. In 25+ years and numerous crimes (thefts, assaults, trespasses, burglaries, hit-and-run car incidents etc.), there was only ever one conviction based even vaguely on actual footage... and that was of a teacher in a school corridor breaking up a fight. Most footage is useless precisely because even with the best camera you'll be lucky to get more than "a guy in a black hood, with black clothing and black shoes hid his face from the camera in the dark".

But my experiences with all that told me many years ago when they first came in that wifi-only cameras were the dumbest idea in existence. First, they're a nightmare to reliably power (batteries running out, not enough solar, etc.) but second they can be jammed by the cheapest piece of hardware ever from quite a distance.

And CCTV in general is only useful is someone is AWARE of what it's picking up as it happens. After-the-event recordings are only for evidential requirements, not crime-detection. If you don't have someone watching, being alerted, being responsible for acting upon the footage (like you would do with your own house if the CCTV pings your phone to let you know something's moving indoors), then you're wasting your time. All you've done is record your security weaknesses for posterity.