'engineer'
'engineer'
682 publicly visible posts • joined 23 May 2022
This is pure Microslop - and if somehow you missed it, Github is Microslop. Enshittify everything with worst in class AI, inevitably suffer reliability going to hell and other systematic failures, then claim you're 'taking this all very srsly' and 'we feel your pain!' and 'we will do better from now on!' and of course all of that is a lie. Because the only thing they care about is AI, and everything else HAS to suffer for that.
Sure glad I got off the shitter for a better host (codeberg). There's no excuse to be on github any more unless you're stuck there because your employers are stupid.
Great, put two of Microslop's worst in class products together and you get... super extra worst in class?
Well, there's still Grok but at least that's superior for generating terabytes of CSAM since that's what it was designed for.
Modi is quite the Trumpian tool (as in he is an angry, racist, malevolent jerk, not that he's the Orange Angry Toddler's tool) and could well have posted something nasty.
And then someone just doing their job who didn't know he's a powerful autocrat who they're supposed to look the other way for took it down. Whoops.
Well who knows. Facebook is just as evil as he is, so who knows who was more evil this time.
I've dealt with Brinks employees before and they were total clowns.
As usual, being one of the best known just means you have the best schmoozing and most sociopathic execs who are (probably) buying potential and current customer company execs the most drugs and booze and whores, not that you're any good at what you're doing.
It's because Slopya put literally all your eggs in one basket, AI. Their evil faceholes spout a lot of lies about how they care about reliability, customers, usability, etc. etc. by they demonstrably do not. All they care about is AI. And then they fumbled the basket with worst in class AI (okay shared with Grok), CoPilot and let that piece of shite write all their software and manage keeping Azure 'up'.
And since XBox *wasn't* AI, they didn't know what to do with it and just didn't even bother managing it with predictable results.
We've always been told executives and C-levels are super smart and talented because this is a Meritocracy (TM) and that's how they got to the top, and of course they're always telling you how brilliant they are.
Well, first we had FElon destroying that myth. The more you found out about him (because he couldn't shut up or stop doing stupid things) the more you realized he was just a moron with a talent for investing his daddy's money based on what smart (if evil) people like Peter Thiel told him to and getting people to make things happen for him. Two extremely important skills for an exec. But technically and otherwise he's as dumb as a rock and emotionally he's an 11 yo edgelord screaming stuff about your mother on a Call of Duty voice chat. A far cry from his previous 'most brilliant man on Earth' facade.
And now, every single executive at every single one of these companies that just went in gung ho feet first for AI with zero concern for consequences or cost is exposed as completely incompetent. They leaped into this because they were on the email list run by the AI companies which told them 'hey this is the hot new trend this year!' and they believed it (and for some reason Desi execs seem to be on another dozen mailing lists pumping that garbage. ) But they're all dumbasses whose only talent is schmoozing. Of course they should all be fired for gross incompetence, but all the board members are also their fellow dumbasses precisely so none of them will ever get fired for being dumbasses. But at least it's all out in the open, there's no defending that 'Meritocracy' any more.
So instead of fixing their completely broken software dev and QA process and maybe not just randomly releasing vibe coded (and vibe tested) shite, the answer is to install a splash-guard for their fire hose of fecal matter. Yep, this is totally on brand for Microslop.
So what are the chances that buggy python code was spit out by CoPilot? Pretty high, I'd guess, especially since they're claiming v2 was partly co-designed (lol) by CoPilot. And nobody on the team actually has any idea what's actually going on, physically, electrically, quantum-ly, or otherwise, but as long as Microslop is funding them who cares?
But pretty obvious now this is never gonna work. It was all completely hand wavey with only a single 'working' sample and everything was based on the python code finding indirect signals in the data, nowhere near what other qbit tech has demonstrated, which... oops. So basically they have vibe proof (in all senses of 'vibe') that it works. 'You just gotta beweeeeb!'
The biggest lesson any smart person would take from Windows 11 is that vibe coding suxass and makes your code flaky, fragile, bloated, and unmaintainable. It's just Windows 10 with half a decade of slop crapped out on top of it.
Of course this has happened with every single product they have. But also of course they won't learn this lesson, can't learn this lesson, as long as Slopya is in charge and every Microslop employee is rated mostly on how much stinking slop they can squat out on their customers' heads.
All these things which have worked just fine for literally over a decade, like the media creation tool which also broke recently. There is absolutely no need to touch them. But then CoPilot goes 'oh hey i have a great idea' and just checks random code in and Microslop runs it by QA, which is also CoPilot, and CoPilot says 'sure that slop looks good' and Microslop ships it because that's what Slopya wants.
This is what full on vibe coding looks like, our glorious future.
> Short circuiting that journey with AI tools makes folks for whom the journey is the reward feel cheated.
Yes, I have experimented with coding LLMs. Gotta keep up to date, it's interesting, and I have to know what the monkeys are using so I can discuss it with the right ooks and hoots. That's the personal thing you're talking about.
But the giant problem here is that those of us who care that things work, are secure, and are maintainable are now surrounded by monkeys who have been given dynamite by slack-jawed executive twits who know nothing about the technology. And the monkeys are flinging the lit dynamite mixed with their poo into the production codebases that we have to work with. This is far beyond personal choice.
Let me change the analogy: The monkeys are flooding our codebases with a firehose of poorly thought out crap that only works on the single use case they told Claude about and tested it on. They also had Claude write the unit tests, which are just laughably wrong and insufficient. And Claude (and the others) can do little apps / utilities, but not possibly keep track of a large / mature codebase - they don't have context windows / sliding attention windows anywhere near large enough to hold the details. They tend to get confused about 2 calls away and multi-threading is just ohmygawd. But hey, passed the unit test and my one use case, merge it into main!
Oh, but that's just bad process you say, it should be in a branch that gets properly reviewed and pulled into main... but the monkeys are spewing so much crap into the repository in an effort to seem productive that nobody can possibly keep up with code reviewing it. But all the herp derp managers see is checkins went way up! Wow, such issues 'resolved'! And since the AI said so and the AI is never wrong, we can just check this all into main.
Going back to your pre-git version control (rcs and cvs and svn represent!), this AI rush is like executives going 'version control is slowing us down too much, let's just all edit the codebase live in a network folder. We'll call it vibe version control!' And you get about the same results.
Of course your point was 'well, the devs will eventually whip this into shape, like they did with version control.' But this time the adoption is WAY ahead of the technology's maturity curve, because most execs are dumb and gullible (though transcendentally expert schmoozers) and everyone has told them they need to be using AI to make numbers go up. There absolutely are places where 'AI' is very useful - it's awesome for OCR, image upscaling, protein folding, math proofs... It's a tool which is great for certain uses. But in most organizations it's just incompetents eagerly applying their new chainsaw toys to open bottles, pick and place components, calibrate pressure transducers, and trim pubes. And at some point the non-subsidized cost of tokens and the new mountains of technical debt building up right now will come due. You only have to look at Microsoft, where absolutely every single thing they do has gotten noticeably worse in the last couple years, to see what it looks like right now when you vibe code everything today.
> Microsoft is facing AI-related issues on multiple fronts. Disgruntled investors have flung a sueball at the company over its Copilot claims,
What's missing here is that EVERY SINGLE ONE of Microslop's issues at this moment is AI-related, with Slopya 100% committed to drowning every single one of the entire company's products in AI generated fecal matter. And that is literally all he and top management care about (they claim otherwise every couple months for PR purposes, but that is demonstrably a complete lie).
So of course security, reliability, usability, everything else goes down the tubes. 100% predictable and predicted.
Indeed (I worked on it with at least one product), so I decided to look it up, and in typical Microsoft (this was before Microslop) fashion there are at least six embedded versions of it (thanks Wiki):
- Windows XP for Embedded Systems - bog standard XP, with embedded only license
- Windows XP Embedded (XPe) - this is what you're probably thinking of, and most people think of. It's what I worked with. It's componentized XP so you can only include what you need. It was used in clocks, slot machines, cash registers, ATMs, arcade games, NAS, navigation systems (like TomTom), and yes, railroads.
- Windows Embedded for Point of Service (WEPOS) - focused on cash registers, ATMs, gas pumps, etc.
- Windows Embedded Standard 2009 - Basically updated XPe with new stuff like .NET 3.5, WMP 11, IE7, etc.
- Windows Embedded POSReady 2009 - Similarly updated WEPOS.
- Windows Fundamentals for Legacy PCs - thin XP variant for dinosaur PCs (back to Pentium 1). Okay technically this is not embedded, but the hardware's close enough.
These were all eventually succeeded by Windows 7 versions and then Windows 10 versions. For obvious reasons of it being a flaky piece of shite there are no Windows 11 versions.
Azure is worst in class (like everything else Microslop does) and falls over left and right (that hooker joke is too tired). People use it because it's shitty but cheap (at first) and bundled with the rest of your Office shite. Not because it's any good. Why would you possibly expect that moving to it would *improve* things?
> Darcy said that the SAM Bus is terribly designed. "There is no way to see the current value without scanning the bus," he said. "But scanning the bus kills the unit."
So you're saying... this is a Microslop product?
Very related, it's pretty funny that Copilot inadvertently exposed the problem by just randomly flailing around as usual.
> after a pilot claimed the AI assistant could claw back 43 minutes a day from administrative work.
Well yes, if you don't care if the output is correct at all you can save a lot of time. Thank heavens the NHS does not!
Which is further made obvious by their picking worst in class AI Copilot* for this.
* Okay, it shares that throne with Grok
This seems counterproductive. Why are you shielding management from their complete f@#$ing stupidity?
If anything you should be tripling down on toxenmaxxing. Give those stupidass ignorant sociopaths what they wanted, which is $500M in AI bills in a month. That will surely make your shows better.
> Have Microsoft never heard of QA testing, or checking that stuff still works in a UAT environment before integrating whatever badly thought-out AI hallucination they have come up with today? People's businesses, jobs, and livelihoods depend on basic things like their git repositories behaving as expected.
All their QA is AI now too. And Slopya doesn't give one single f@#$ for your business, job, or livelihood. He knows you're stuck with Microslop because they've made a 'great deal' for flaky shite like Github / Office / Teams / Azure / etc. with your dumbass executives, so nyah nyah nyah.
(Seriously, I feel you, but they do NOT CARE. And never will as long AI is the only thing Slopya cares about.)
Given Google this was probably a totally automated AI decision. Or some dumbass support tech (#notallsupporttechs) deleted the wrong customer.
And then when you make a level 999 support call, The Chocolate Factory is all 'oh our AI wouldn't make a mistake' and takes an hour for anyone sentient to even notice.
Google I/O is coming right up and they're already trumpeting about how deeply AI will be integrated into the next versions of Android. Like agents ('robots') offering to do your shopping across multiple apps and lots of other 'helpful' things, 'improved' AI autofill, Gemini even more aggressive about offering to 'help' you... yes, baked into every bit of the system. It sounds even more annoying than Copilot, which is pretty amazing.
Soooo... I moved to iOS on phone a while ago because of Google aggressively enshittifying everything, and it's much better (once you turn off that Liquid Glass shite, everything has annoyances). Very little AI and you can turn it off. Might be time for that unless you are up for completely going to GrapheneOS. For now I still have Android on my tablets so will be watching this omnishambles with a front row seat.
I just moved to GitLab to get away from GitHub's relentless AI enshittifications.
Is there any place safe (for now)? Yes, I do just keep most of my git stuff on my own servers but there are things I need public repositories of with a little bit of interaction with other contributors.
If you're still using Office365 it's only getting worse from here. It can't possibly get better as long as AI Uber Alles is their corporate mission.
I bought a perpetual license from a key vendor for Office 2021 (I forget exactly which one) for $40. It's beautiful. It works fantastic, no monthly fee, NO AI EVER. It is safe from Microslop's relentless enshittification of everything.
Now apparently these keys are a bit shady, but it works, and f*ck Microslop. I'm only doing this because I need Excel to collaborate with outside engineers (and Libre isn't compatible enough) and then you made Office so shitty I couldn't stand to keep using it.
Of course, if your office is locked into Office 365 because your C***s made some shitty 'deal' with Microslop, then you just have to suck it up unless you have enough permission to run Shut Up Windows or RemoveWindowsAI.
> Downvoted for the use of the word "literally".
I MEANT literally. They would /literally/ be throwing live kittens and puppies or toddlers (okay, having underpaid workers do it) into giant grinders if they could figure out a way to make enough money on it. That 'literally' is there to make it clear that they have no ethical or moral limits and would actually do so. Why would you think MBAs would not?
If you're working for some shitty corporate hellhole like Oracle, Microslop, SailPoint, Andersen Consulting, IBM, Boeing, Sage, Capco, Intel, Nvidia etc. etc. etc. who are MBA driven and actively hate their employees, why wouldn't you sell your work credentials? Why would you have any loyalty at all? The corporation certainly doesn't deserve it. All the above listed companies would literally grind live kittens and puppies into blood meal if they could make a decent profit off it. They would grind YOU into blood meal if they could make a profit off it. So why wouldn't you sell your creds for a good offer? It's what an MBA would do!
> Git was designed from the start as a distributed system. I'm probably missing something, but the whole idea of a centralized repository seems like missing the point entirely.
Yes, any git repository is the entire repository. So the one on my laptop is just as complete as the one on Github or Gitlab. But like @Richard 12 said, the one you point people at is 'the' repository.
And with Github/Gitlab it's also where you point to downloads, have the bug tracking, make announcements, etc. etc.
People are just not wired for complete decentralization, they want a place to go.
> Or opt-in to having your use of Copilot announced to the world, making keeping schtum about your little habit the default?
The biggest(?) problem here is that it was telling everyone you were a shitty vibe coder using Copilot *even if you weren't* - even if you hated everything it stands for. You are now, for eternity, 'on record' as being one of those chucklef@#$s even though you were completely innocent.