Adverts and telemetry
How else do they get enough compute power to serve all the pointless adverts and collect all that juicy telemetry.
1715 publicly visible posts • joined 11 Dec 2014
I came here to say the same. I recall using my old Psion II with an RS232 interface to talk to a Plessey PBX sometime back in the mid 80's, it was easier to do that as laptops weren't really a thing there and I didn't have time until later to run a proper RS232 cable to my desk, so I could handle it better next time.
I've still got the organiser somewhere in my long term storage / thermal insulation layer - the roof space. I bet it still works if I put fresh batteries into it.
At least we can trust the information on LinkedIn to be 100% accurate, so there is nothing to worry about.
I'm really pleased to see that virtually every profile is for someone who is the founder of some single person company, or some over inflated and completely meaningless title, or that they are the CTO or managing director etc.
Perhaps now is the time to go back to polluting the data set with the "Advanced Llama Trainer" skill and other such meaningless names. Lets see how the analytics copes with that and how the Government decide which jobs they want people to skill up for - because stats.
On the flip side, lets stop offshoring everything, so that there are actually jobs that people want and can mature into - just like it used to work before.
You can shorten that to "unnecessary copilot"
I documented a build for a Win11 system the other week, the amount of pointless screens and endless faffing about was huge. Making notepad a bit safer, then installing notepad++ was one area. Another was the sheer number of prompts to just get Edge to display its first web page. There were about 20 different questions and prompts.
Back in the old days, you could get the OS up and going with a lot less pointless screens and dialogue boxes.
So, When was the last time Linux had such an issue. Then, when was the last time that a Linux user couldn't work around it by removing a package or making a config change ?
If you want double-click to do candy crush, fill your boots.
If you want a proper operating system, then get learning. You might be pleasantly surprised with what you find out.
Silicon yields are always important. Flipping a couple of fusable links to permanently turn off the bad bits is always an option.
This way, partly working devices can be sold as functional devices, with a lower spec. I see nothing wrong with this approach, since it keeps things out of the rubbish bin for a bit and cost / performance is always a sliding scale. How much CPU power / RAM capacity / cost is enough for your project ?
Probably because it ate its MicroSD card.
This is generally when I decide its time to rebuild mine, then its always a rummage in the spare parts box to see what might be a better solution for the next couple of years. Then when I've realise there is nothing useful in there, in comes the next delivery from your favorite delivery company of the day and off we all go again for another couple of years.
If only they would test their own software properly before rushing it out the door, then a lot of this could be contained within their offices, rather than crapping on their customers all the time.
They clearly are not learning from their mistakes, or perhaps they just dont care.
My point is that MIDI is a stupidly simple protocol that has been around for donkeys years (and no el reg doesn't have a reg-approved unit for that).
But, if you look at all other similar low baud rate serial protocols that have been around for decades and take very little processing to make them work. Most have similar functions within microcontrollers and microprocessors. So, needing to get Windows 11 to finally provide MIDI2 support seem to me to be stupid, in the same way that if the article was saying that Windows 11 now supports other such protocols - I2C, SPI, UART (RS232, RS485, RS422, etc), 1 wire, CAN, then people would be laughing.
Its not difficult, so why has it taken multiple decades to get Windows support for this ?
I accept that the BBC micro may not be the platform of choice now, it was an indicator that any old CPU can handle the workload.
Of all the things they could have done - Why now
Why after all these years, surely most aren't playing with MIDI now, or if they are, then they are using something a bit more reliable than Windows, with its forced updates. Imagine that in the middle of a concert.
Why is there not something more important they could have fixed with the same engineering resources, such as notepad.
Perhaps they are training their new vibe coding AI, v14.3 to see if it can do basic stuff.
Even a 1MHz BBC micro from the 1980's could do MIDI reliably, with only 16K of ROM and 32K of RAM. It wasn't a difficult thing to do even back then - 40 odd years ago.
I've always seen it as a fuck design, do stuff and look busy, then say that any bugs are to be fixed in the next sprint.
What a way to do things. We used to call it bad practice back in the day. You know, decide what we wanted, do stuff that deliered that stuff, management were generally happy.
The good news though is that what goes around, comes around, so we are now closer to doing it properly again.
I came here to say the same.
You will know which one you have in a couple of years time. If its the old one that just keeps going, then its the original, good recipie. If however you have forgotten about it and replaced it several times, when it randomly went ping and wrecked your meal idea for the night, then you had the new cheaper one. Just remember to say thanks to the accountant that changed the recipie as it saved them a few pennies and wrecked the reputation in the process.
I believe you can tell the difference by the hue across the glass when you look at it a bit percularly. There are many articles on line about how to do this.
One area where this will clearly be applied is in the first response to all customer trouble tickets, with an AI slop type answer of course.
By pure coincidence this means that the metric of first response time will drop to zero seconds, meaning a 100% success metric on that parameter of any SLA.
Rinse and repeat on things like first line and second line. Customers will now be fighting AI, which will keep changing its mind and hallucinating, but SLA’s remain safe, so no service credits for you either win-win for everyone except the end customer.
I was poking around in Notepad the other day, as I had to use it as NP++ wasn't installed. I found that there are some settings now under settings (cog icon), top right, which allows some of the crap to be turned off, including Word wrap, formatting, not re-opening everything when you re-open the apps, discarding previous session, no spellcheck, no auto correct, no copilot.
Each setting is obviously defaulted to on, which is generally not what you want. Its all very much Wordpad lite now.
I'm also left wondering who the target user base would be for this monstrosity - given that they are driving everyone towards Office 345 and anyone with half a brain will be installing one of the Open Source alternatives if they actually want to write docs for free.
Bring back notepad. Simple, like it should be.
I wonder if Microsoft could publish a simple paper about what their expected use cases are for all the overlapping products that can edit stuff.
Perhaps if they actually delivered the next generation service, so that, well, customers could actually order the new services, then their uptake may be a bit higher.
For me, fibre has been promised for over a year and the status tracker shows that its being delivered in our area, yet oddly, we never see anyone with a van, nor any of the manholes up, or people up the poles. The only sign of life was that someone that could operate a hammer, nailed a sign onto the pole saying that there is overhead fibre, when in fact there is nothing new on the pole as the trees are still right around the pole and there is zero sign of anything modern at the top of it.
I'm also wondering what happens when they threaten to turn off POTS, realise that not that many people actually care about the old telephone network as everyone that can, has already moved to mobile devices. The good news will be that there will be less calls fraudulent / scamming calls
Whats wrong with rolling back, when things go wrong - Both Microsoft and allowing users to do the same, then the rush won't be as bad to fix things.
The number of these sort of problems thoguh is really worrying, they clearly don't give a rats ass about how we might use the computers or try and run our businesses on them.
I'd be interested to know if this is AI related / Vibe coded, since its a useful yard stick to see how bad things are.
It would be greate to see this sort of tool merged into WinAeroTweaker, so that all the settings can be done in one go.
On the flip side though, the more tools that are created that push back to Microsoft to say "no thanks", the better.
Remeber folks, its your computer, not theirs. If its their computer, then they would be paying to fix it each time it breaks.
So, one rule for them and one rule for us ?
This is from the same group that want to have a National Identity database, with all our information.
Arguably, they should have MORE stringent rules applied, to ensure that things are appropriately designed, governed, implemented and operated.
What happens to the outsourced companies that deliver these services ? Are they to comply with the "for the rest of us" rule, or the "don't worry, were different, government" rule ?
Whats surprising to me is that nobody has mentioned that V6 addresses are well known to be 128 bits, which is also 2 x 64 bits, so with modern CPU's and with some great foresight on network addressing and segmentation around a /64, thats exactly one instruction to compare addresses, in a CPU (or other hardware such as FPGA's), meaning that routing, at the core of the Internet, on the Tier 1 ISP routers, becomes much simpler which in turn means much less computational overhead and much less RAM requirements - as the other 64 bits are irellevant for routing. This has the big benefit of faster routing, which means lower latency for everyone as their packets arrive faster.
There is the benefit for all users, since the local LAN part is generally a /64, then the same sort of higher speed routing or switching can be done, again improving performance at customer sites, where hardware will generally be lower power, compared to service provider kit.
Obviously the middle layer of routers, between Tier1 and customers, where masks will be variable, but generally on /48 or /56 or similar. will need to handle the whole address, but thats the price of flexibility and exactly as it is in v4 (through the whole stack, Tier1 to customer). Now though, the intermediate routers won't have to worry about upstream or downstream, again allowing routing information to be stored more efficiently in the hardware and again reducing routing table sizes.
Its also clear from the posts, that there are two camps on v4/v6, mostly around those that don't understand it well enough vs those that understand it better.
All the "invent a new protocol called Vx.y, where things are magically fixed" always glosses over the massive problem that you can't put more than 32 bits in a 32 bit address, without breaking everything, all software of the day allocated exactly 32 bits for storage of addresses, so any hypothetical expansion (irrespective of what you call it) will need a bigger buffer in RAM on every app that use IP addresses - meaning ALL network enabled applications on ALL devices will need their software (hardware for network devices) updating to have the larger buffer size and operate correctly on it, then recompiling, testing and re-installing.
Untul this is "fixed" in some magical way, across all global devices, then nothing would work. It should be crystal clear that this is not a viable way to "fix" things, so the v6 approach of parallel run removed that massive hurdle, allowing soft take up and transition, which is still happening, even with people trying to deny it.
@Nanashi clearly understands his stuff, so thanks for the high quality and accurate posts.
Perhaps what actually happened is that the wolf cub decided that the Microsoft marketing hurt its eyes, or perhaps there was a Douglas Adams moment - ‘ oh no, not again’, given the prospect of having to use Microsoft software again.
Well done Wolftone and Mrs Wolftone for the prompt introduction to IT at an early age and congrats on the new arrival.
Well, thats going to make it more interesting when they mark any assignments then. What sort of increase will they get in people that will use AI to generate their coursework ?
Kind'a undermines the whole idea of critical thinking and learning, if they are taught how to use AI to not have to think.
I got into trouble, for suggesting to a user that, perhaps, they should write down the instructions. This was because I had had the same issue logged for 3 consecutive days.
I think it was "unable to access e-mail", but in their defence, it was back in the Windows 3.1 days, and the users had two mailboxes, one on an IBM mainframe, and another completely separate mailbox on a VMS system.
So, access was simple, you were getting used to this Windows 3.1 thingy an you had LAT running, so DECNET access was in place, but needed to manually start the Trumpet Winsock IP stack, then open the 3270 terminal emulator (Rumba) for the mainframe, or open WRQ Reflection for the VMX box. Once in the app, you then open the preconfigured session to access the remote host, then login to the server, then go to the mail app, then navigate the mail app to read or write your emails. Once you were done, follow the logout procedure.
Yes it was stupidly complex for ordinary users and a right pain in the a$$ to make everything work on the systems, but it worked and it had been rolled out by the powers that be. It then fell to the army of support people to keep it all up and running.
What I didn't understand as a young 20 something at the time, was that there was this thing called menopause and the meer act of suggesting that perhaps writing down the information would be a good idea, got a complaint filed against me. Luckily management saw the bigger picture and the lady in question actually apollogised to me a couple of weeks later.
The support days are good fun at the time, but you are sure happy to have them behind you, once you work your way up the ladder a bit.
So, the government claim is that people are "Baffled by AI" - er, no. It' that most people don't need nor want it, thats a very different thing.
Then there is the small cost of living thing - people are putting their hard earned cash into things they actually need and want - little things such as food, constantly rising bills and paying the mortgage or rent. Other things are pushed down the pile.
AI is pointless against all of this, each and every one of us has Real Intelligence - the stuff that is in our brains. Sure there are different levels of Intelligence, but thats the same with AI, except that people tend to hallucinate less and its vitally important that people are taught to think, since that how we evolve knowledge. AI can't do that, it just sucks it all up and regurgitates it - without any quality filter of good information vs poor information.
I'm also left pondering what a 2025 reinvention of the BBC for "technology skills" would actually look like - sticking an AI capable system in everyone's house is a bit different from a single board 6502 based system, plus there is the Elephant in the room - the lack of availability of components, since AI is sucking them all up.
“We’ve read your posts and heard your feedback,” GitHub said. “We’re postponing the announced billing change for self-hosted GitHub Actions to take time to re-evaluate our approach.”
So, they have not cancelled the idea, just postponed it, which means it will rear its ugly head with slighly more glitter and marketing BS before they try and force it down your throat again.
Vendors need to understand the difference between "mine" and "yours"
Luckily, I've only ever had to touch the platform layer for ERP systems a couple of times during my career and in most cases, the bad memories are sufficiently faded.
But, I distinctly remember a previous customer telling me that every so often, they decide to buy a new ERP platform that doesn't do what they need, then they spend the next 10 years customising it to fit the evolving business needs, so that in the end, the only thing left is the product name. This is apparently cheaper, easier and more likely to get board approval, than trying to steer the battleship back onto course to do things in a more industry standard manner, where new staff could be hired in that know the process, rather than having to be told "we don't do it like that here".
Now, I'm no accountant, but I assume that the basis of such systems must be broadly similar across the world, with the need to count materials and beans coming in and then product and beans going out, along with a shake of regional laws that don't line up with each other.
So, surely there must be an opportunity to produce an open source product that could be used by enterprises, with licencing such that they must feed back re-usable (ie industry wide) customisations back into the product. This way, there would be a product that does much more of what people need.
I assume that much of the change comes from legal / regulatory (tax) changes over time and common business events such as mergers and acquisitions.
It could be architected using a modern cloud native, scale-out approach, rather than the old school monolithic behemoth, making it easier to platform on different cloud providers and scale for different business sizes. This would also help when people decide to re-baseline, since modules can be considered independantly to the thing, making updates easier.
Since the government is very fond of rating systems, such as those used to rate schools. How about if they implemented the same thing for the outsourcing companies that they use across all government programmes. Then they could try and work out which ones are no good, so that they don't keep making the same mistake.
The trick is that they need to consider cost, risk and time. Just because it was cheapest, doesn't mean that they will get it right.
You would also assume that with a win on a £239M cointract, they could afford at least a handful of competent people to drive it forwards properly
Given the number of things that have been reported to be visible and wrong, you can be fairly confident that the same sort of level of quality will exist all through the platform.
Try when you are talking to an offshore resourcing team and they have really lousy Internet connectivity.
India, I'm looking at you.
We had more success listening to the birds in the trees than the person that was presenting the latest design for approval
Have we lost Arav again ..
Yep, OK, lets wait for him to rejoin.
Wait ...
Do I have audio ?
Yes, you have Audio.
2 more points on the design.
Then, again, we are waiting on them to re-join.
Good people that are knee capped by their poor connectivity
Looks to me like we are all using Windows Yee-Haw edition.
You might get lucky today and it doesn't kill you, or you might not.
FFS MS, use some form of testing on your software.
Perhaps even train your AI to do something such as basic testing and prove its not completely pointless.
It doesn't say when that happens. Is it on submission - as you would hope, or is it some form of asynchronous background process that gets done when someone gets around to it and as we know, people have plenty of time, so there will be absolutely no delay, promise.
Or, with AI everywhere, perhaps, it is some automated code that can't figure things out properly, just approving changes and firing them out there. AI has good form for this, it can simply lie or say sorry when it gets called out that it made a mistake.