Microsoft goes native with Copilot. Again Microsoft has revamped its Copilot app on Windows once again, this time insisting it really has gone native. The update, which is rolling out to Windows Insiders from the Microsoft Store is, according to the IT giant, a native XAML app. It includes a side panel for starting conversations and a conversation history list. …
"How do I get this piece of cr*p Copilot off of my PC ?"
Funnily enough, that was the first question I asked it when Copilot* appeared on my PC. The answer it gave me wasn't very good though. But having had 24H2 forced on me, it came back again. I was pleasently suprised that the Start menu gave me the option to uninstall it. But I'm sure it'll be back again.
But in my continued attempts at being a garbage collector for MS.. Why can't WIndows Update have an 'advanced' option, like a lot of other software? So give me a choice about the features I want, or don't want installed? I was dreading 24H2 after hearing that it forced HurtLocker onto users, but I managed to avoid that on my and the kid's PCs. I really don't get that decision given MS is basically doing what ransomware gangs do.
Also..
However, it does show the direction of travel – at some point, folks will be able to control more of their PC using a conversational interface rather than a terse call..
I'm.. not convinced. I guess if that means voice control, and it can deal with profanity, it.. might do something useful. But that isn't how users or OS's really work. So I click, or double click on an app and do my thing. Could be fun in something like Word though. Yo, Copilot, write me the next Harry Potter novel.
*Halo appeared on Netflix and it's been amusing watching the product placement for Cortana.. the 'AI assistant' for poor'ol Master Chief who just seems to be annoying and a hinderance rather than anything helpful. Hopefully Copilot goes the same way and gets discontinued.
Once upon a time I'm sure readers remember the 'Uninstall/Install Windows Features' back when you could decide what weird and wonderful stuff you wanted on your Windows PC (see icon). That seems to have gone by the by.
Apple are a bit better here. Apple Intelligence turned up in one of the latest MacOS updates BUT it didn't enable itself until you told it to, and it's dead easy to turn it off again in the System Settings.
Once upon a time I'm sure readers remember the 'Uninstall/Install Windows Features' back when you could decide what weird and wonderful stuff you wanted on your Windows PC (see icon). That seems to have gone by the by.
It's still in there, under programs and features in the control panel (available in the left-hand menu, requires admin privs). At least it is on my home machine.
Admittedly it's buried down deep enough to be recyclable as firelighter, and so deep into the maze that is the current Windows start menu that you risk running into a minotaur if you take a wrong turn...
same as I am doing ahead of the demise of win 10
upgrade windows to Linux :)
1 of my 3 win10 machines was migrated to a raspberry pi4 (was doing simple stuff with receiving weather satellite images)
2 machine to go, linux Mint seem to be my favoured weapon
...the old favourites like "something went wrong"...
Think my all-time favorite was along the lines of 'An unknown application has caused an unknown error'. But I also just love the way 99% of error messages don't seem to let you copypasta the error so you can Bing it in a forlorn attempt to try and figure out what it means. And then Google it because Bing is about as useless as MS Help pages. Not that Google's really any better, but sometimes it might lead to something on Reddit that might help.
Oh, and the 'please wait while we upload the problem details' thing that I presume goes to MS's dev/null somewhere because since that first became a thing all those years ago, I've never seen it provide me with a solution. I guess this might be why I'm just a tad cynical about the supposed benefits of MS AI.
My favourite error message is:
Error: The operation completed successfully (0).
I guess this presumes that the engineer normally expects the operation to fail!
This happens when a programmer blindly treats any erroneous return value from a function as failing due to an OS error and they include the OS error string in it, assuming that contains something useful, but it doesn't half make for a stupid error message.
Help pages being unhelpful ... they're either written by some poor intern or by some prototype AI. Either way up it's usually straight off to YouTube to find people who actually know this stuff.
While I'm having this rant, why does the GUI in Exchange Online Admin not work when setting up rooms and resources? Why then does the help pages ask you to install Exchange Online PowerShell and type all kinds of mysticism to set this up? Did the writers of the help pages KNOW that the GUI doesn't work? And the Room Finder doesn't work because it expects a Building to be selected first when there is no way to input a Building in the Exchange Online Admin GUI? And this passed testing??? (only joking, we all know MS doesn't do testing, just heave out an update and let the feedback commence).
if the AI was up to snuff, it would know that something amiss and fix the problem with the printer without your having to engage in a conversation with Clippy++.
But as anyone who has ever had to Google for a solution to a Windows problem knows, the only available training data for AI that might be deployed in these circumstances consists of thousands of almost identical web pages suggesting that you turn everything off and on again and then install new drivers and "registry cleaners" from a variety of dubious sources. I can't see Microsoft investing the time and effort to produce reliable and comprehensive troubleshooting documentation because it would imply they had the knowledge as well as the time and effort to make their software more robust, so where is the AI-enhanced advice to come from?
And, in any case, offering AI as a contaminated sticking plaster to cover already suppurating wounds sounds like an act of desperation.
Of if from the MS Community, some plank who has no idea who will start of with something like
"HI I am clippy,
I have 12 years working with Microsoft Windows and I am here to help you
Please reboot / try a full install / sorry, Windows cannot do that"
The last response maybe after a few weeks of telling Clippy they are a twat, we have tried everything else and it still does not work
Simple query, "how do I get my printer working?". What then happens?
I assume that the system looks at your machine, takes a full hardware and software inventory, uploads it to the system where your data is absorbed into the Artificial Anus as training data so MS can make more money out of it, then suggests the user should watch a couple of episodes of the IT Crowd to learn how to reset their system?
I don't want my system data uploaded - if details of system configurations are made available to all and sundry beyond my firewall (even "accidentally") it could display an attack surface or provide data to advertisers.
So how do I stop that trawl happening? Does GDPR "legitimate interest" not include the right to refuse to have data uploaded unless the user consents to it - assuming if it is personal and, in my opinion, software configurations of machines could be very personal - eg finding drivers or hardware identifiers for Blow-Up Brenda and Vibrating Vicky?
"Good question. Before windows 8, setting up a device in Windows was straightforward. You would go into Control Panel and find the section you want. Since 8, thanks to constant UI changes and settings having a split personality disorder, I'm fucked if I know. I hope that answers your question".
... given the Old Microsoft "Helicopter Joke" from the mid 90's:
"Whilst flying around Seattle on a very cloudy and hazy day, a helicopter suffered an electrical malfunction that disabled all of it's navigation and communication systems.
"Due to the weather conditions the Pilot couldn't determine their exact position or the course to set for the airport.
"The Pilot flew to a tall building and, whilst circling it, wrote a sign asking 'WHERE AM I?' in large letters and showed it to the people he could see in the building.
"The people in the building quickly responded with a sign that read 'YOU ARE IN A HELICOPTER.'
"The Pilot smiled, waved a thank-you, looked at a map and promptly set course to Seattle airport where the helicopter was safely landed.
"After they were on the ground, the Co-Pilot asked the Pilot how the 'YOU ARE IN A HELICOPTER' sign helped determine their position.
"The Pilot responded, 'I knew that had to be the MICROSOFT building because they gave me a technically correct but completely useless reply'"
= = = = =
The Pilot figured out the solution, the Co-Pilot couldn't...
In all honesty it would be useful if it could do stuff that really did make things easier for the user, and more importantly, if it’s reliable.
What’s the end goal?
GUIs promised to hide the command line interface from the user. AI seems to promise to remove the GUI for the user too. We’ll end up with computers that actually do the stuff you need - based on a required end result, rather than individual commands.
The future is bright, but I really don’t want Microsoft to be the one attempting to deliver it. Like everything else they do, it’ll be obfuscated, awkward and restricted by a myriad of different subscription options. In other words, shit.
I had an Ubuntu 24.04 installation for a few weeks, but Friday I began the switch to LMDE (Linux Mint Debian Edition.) Took no time to set up everything on it - CUDA stack, Ollama. AnythingLLM, Zonos, Java stacks, Python stacks, Docker with CUDA enabled, etc.
The only issue I had was caused by a recent VirtualBox update - 7.1.6 won't run Ubuntu 20.04 or 18.04 VMs, which I need, so rolling that back and freezing 7.1.4 was fun.
(Not my downvote).
Don't know if it's related, but I also have VirtualBox installed on a Mint host (regular, not LMDE), and I have various VMs, including a Mint VM I use for testing things on first.
The Mint VM has worked perfectly for about 2 years now, but a recent Kernel update (from 6.8 to 6.11) broke things. After updating to 6.11, a boot would either hang, or after several minutes I'd start getting lots of time-out errors in the boot up log. Managed to get into the Desktop once, but it just hung instantly, after that never got to the Desktop again, just the time-outs.
Tried investigating, but couldn't find the root issue. So went back to an earlier snapshot (I always take one for any major changes), so back to 6.8, and it boots up fine again. Installed all updated other than the 6.11 kernel, and still working fine.
No idea if this is just a VirtualBox issue, but for now I'm not risking going to 6.11 on the host machine either any time soon.
Will likely try a 6.11 point release in a few weeks, or wait for 6.12 or 6.13 to drop (in the Mint repo), then try the VM again.
Just to add to this, in case anyone sees this in the future.
Turns out it was VirtualBox was at fault, and having the issue with 6.11.
Installing the same Mint 22.1 under KVM on the same host, and that works fine with the newer 6.11 kernel.
The newer Kernel also worked fine for the host machine (same Mint 22.1), although I've since switched to 6.13 on the host, as I got new 9070 XT which needed 6.13 as a minimum.
So I'm now migrating from VirtualBox to KVM!
> at some point, folks will be able to control more of their PC using a conversational interface rather
> than a terse call to a family member or IT support person to find out why the printer has stopped
> working again.
I would bet you 100 quid that Copilot will be as useless for actually fixing any problem as Microsoft's websites are. It is so f@#$ing stupid it can't even answer much simpler and straightforward things. If they're calling me it is far beyond Copilot's capabilities.
'Help my sound is not working.' 'Hi, it seems like your audio is not working and you want to know how to fix it, is that correct? Thinking... Hmmm, it looks like your first step should be to pour bleach into your earholes. Please let me know if you can hear your audio after that, and if not I will try to help further.' 'I still can't hear anything.' 'Did you pour bleach in your ears?' 'No, that sounds dangerous.' 'I am sorry, but if you do not pour bleach into your ears I can not continue to help you further until you are doing the needful. But did you know you can use Copilot to generate pictures and summarize documents? Would you like me to open Word and Paint for you?'
Gone troppo or feral might be more accurately and incidentally possibly less derogatory.
Actually Microsoft's frantic embrace of this nonsense ought to be enthusiastically encouraged in the hope that it expedites its ultimate self destruction.
I am always happy to hold the coat of a chap who is about to jump off the bridge.
>> at some point, folks will be able to control more of their PC using a conversational interface <<
Who the fuck is going to talk to their PC in a busy office? Or perhaps, show you are talking to yourself or the wall when WFH? Idiots. It didn't work last time and it ain't going to happen this time.
AI help when searching for a solution to a computer problem is rather like having an enthúsiastic but fundamentally stupid co-worker at your elbow. He's constantly offering useless advice resulting from an incomplete appreciation of the problem and superficial knowledge gleaned from skimming Microsoft's useless help pages.
It's a perfect Dunning-Kruger simulation. General Artificial Stupidity (GAS).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
