back
350 comments
So, a couple years ago Microsoft was the first large, public-facing software organization to make LLM-assisted coding a big part of their production. If LLM's really delivered 10x productivity improvements, as claimed by some, then we should by now be seeing an explosion of productivity out of Microsoft. It's been a couple years, so if it really helps then we should see it by now.

So, either LLM-assisted coding is not delivering the benefits some thought it would, or Microsoft, despite being an early investor in OpenAI, is not using it much internally on things that really matter to them (like Windows). Either way, I'm not impressed.

I know blaming everything on LLMs is in vogue right now; but this is much more to do with Microsoft very publically firing the QA department[0][1] as a cost savings measure and claiming developers will do their own QA (long before LLMs were on the scene). It started in 2014 and the trickle never stopped.

Microsoft has a cultural problem; it went from an "engineers" company to an MBA directed one, trying to maximize short-term shareholder value at the cost of long-term company reputation/growth. It is very common and typical of US Corporate culture today, and catastrophic in the long-run.

[0] https://arstechnica.com/information-technology/2014/08/how-m...

[1] https://www.reuters.com/article/business/microsoft-expected-...

I fully believe highly skilled people can get a great benefit from LLM tools; probably not 10x; but enough that its noticeable.

The key thing for me is that it only works when the LLM is used for tasks below the devs skill level; It can speed up somebody good, but it also makes the output of low-skill devs much harder to deal with. The issues are more subtle, the volume is greater, and there is no human reasoning chain to follow when debugging.

So you combine that with a company that has staff in low skill regions, and uses outsourcing, and while there might be some high skill teams that got a speed up, the org is structured in a way that its irrelevant.

The argument I usually hear is that you only truly get the 10x improvements with <new-model> (right now, Opus 4.5), so they've only had a few months, not years. In a few months, it'll turn out that <new-model> wasn't actually capable of that, but <new-new-model> is, and as that's not been out long its unfair to judge so early. And so the cycle begins anew
Imagine a world where Microsoft was pushing “Copilot” integration everywhere, just as they are in this one—but the proof was, actually, in the pudding. Windows was categorically improving, without regression, with each subsequent update. Long-standing frustrations with the operating system experience were gradually being ironed out. Parts of the system that were slow, frustrating, convoluted, or all three, were being thoughtfully redesigned without breaking backwards compatibility, and we were watching this all unfold in real time, in awe of the power of “AI”, eyes wide with hope for the future of software, and computing in general.

Think of how dramatically this hypothetical alternate reality differs from the one we live in, and then consider just how galling it is that these people have the nerve to piss on our leg and then tell us it's raining. Things are not getting better. This supposedly-magical new technology isn't observably improving things where it matters most—rather, it's demonstrably hastening the decline of the baseline day-to-day software that we depend upon.

They weren't great before LLMs either.

Also, it seems from the outside like a dysfunctional organisation, or at least with incentives heavily misaligned with their users. Replace LLMs with a bunch of 10x engineers and it will still be bad in an environment like this.

So not sure how much to blame the LLMs - or in fact how much MS is really using them. Poor souls have to use MS AI tools, I almost feel sorry for them.

Everyone should read "The Bear Case for AI" thread:

"The bear case for AI is that bringing 10x or 100x or 1000x more intelligence to America will not change anything because U.S. institutions are already designed to ignore or waste intelligence and have no idea what to do with any more of it."

https://twitter.com/mmjukic/status/2014255931215716545

I would like to point out that not even a week ago Satya Nadella stated that someone should finally do something really useful with AI because if no one does then they'll lose the social permission to burn all the energy on training and running the models: https://www.pcgamer.com/software/ai/microsoft-ceo-warns-that...

Mr. Nadella, why not lead by example and make Windows the most amazing operating system ever created with the help of Copilot? What's the holdup?

Oh it did help.

Microsoft went all in on do more with less and fired/reorged significant part of the company.

Wouldn’t be surprised if the outage is caused by new team taking something over with near zero documentation while all the tribal knowledge was torched away

Or LLMs weren’t good enough yet years ago, but the growth curve looked so promising that an investment seemed a good idea.

Also: do you have a reference for “a couple years ago Microsoft [made] LLM-assisted coding a big part of their production”?

I know they started investing, mentioning future benefits, but don’t remember them saying their Windows development team (heavily) relying on it.

I'm wondering why the guy at Microsoft in charge of Windows is still employed.

Over the prior weekend my installation of Playnite (a catalog/launcher for my games) was broken by the update, until I moved its data off of OneDrive[1]. And the other day I figured out that a couple of icons on my desktop had become completely inert and unresponsive due to the same bug - again due to an interaction between the Windows Shell and OneDrive. And this one I can't fix, I can't shift my desktop out of OneDrive.

MS's strategy at this point is that Windows is a loss leader to get people onto the subscriptions for Office and OneDrive. So when the Windows team releases bugs that break usage of those services, forcing people off them onto alternative solutions, the guy in charge of those updates really needs to be answering some tough questions.

[1] I've now got SyncThing handling this.

Unrelated to the article itself, but I'm really not a fan of websites like this one hijacking the back button to show more articles. It feels like a mistake to have ever allowed websites to modify behaviour of the back button like that, although I'm sure it leads to some usability improvements elsewhere.
Here's a similar discussion[0], and here's my experience[1]:

Last Thursday windows 11 forced this update on my Acer machine. It caused me BSOD: inaccessible boot device, so I had to reformat my machine to get Windows running again.

So I am now very wary of this Out of Band Update[2], especially when it's not mentioned whether the latest update solve my issue or not. I don't know the same problem is still there, or whether this update makes the problem any better or worse

[0]: https://news.ycombinator.com/item?id=46761061

[1]: https://news.ycombinator.com/item?id=46761870

[2]: https://news.ycombinator.com/item?id=46750358

How can a company this big fail so hard in what one would consider their main* product still baffles me.

*Yes, they probably make more revenue in Azure or Office365 licenses but at least when I think “Microsoft” I immediately think Windows.

I seriously wonder if everyone in the Windows development team(s) are just vibe-coding everything now. I feel like all of these are rookie mistakes from the POV of working on an operating system. This is also the consequence of eliminating all QA and testing and forcing your users to do that for you. Admittedly there are some things that are hard to test (or impossible to) in an automated way, but that's what the old Windows hardware lab test machines were for.
I see Microslop's "AI" coding mandate is continuing to go well
Just checking into this thread for the record, I commented in one of these "update bricks windows" threads maybe a year ago about how I don't let my Windows 10 update, still haven't updated and everything is still working great!

If your device is working, after an update there are only two options, either it keeps working or it doesn't. Why roll the dice?

That's why I live very well with Win 10 without updates. I downgraded in May last year from Win11 after months of frustration and headaches and since then everything works smoothly and peacefully.
> After the Power-On Self-Test (POST), and you see the blue Windows logo (if applicable), press the power button again to shut down.

> Repeat steps 1 and 2 multiple times to trigger the Windows Recovery Environment (WinRE).

What exactly was wrong with pressing f8?

At this point Microsoft needs to go back to service packs and a three year OS version cycle. Rapid development doesn't seem to be working.
> "Microsoft has received a limited number of reports […]

Interesting working: one night interpret this as “a few reports”, but they’re technically saying “a finite amount of reports”, without really implying if there were a few or many cases.

Open gpedit.msc, configure policies to disable automatic updates. At this point Windows is a virus that is useful for only playing computer games and should be avoided for any other purpose.
Heaven forbid any company ever come to the conclusion that shoving updates down your users' throats against their will might not be the best idea humans ever came up with.
Modern Windows... It's like having your own DoS adversary baked into your PC.
At least they're shipping a million lines of code per month per engineer. That's what counts.
This makes me wonder what the testing method is for this sort of thing. Presumably you'd do as much as possible in VM's to get repeatable states with updates from every conceivable starting point.

But what about physical machines? Exactly how many repeatable update tests on physical machines are done? You get a combinatorial explosion if you need to test on 100 different configurations for hardware and you have 100 different starting points in your pre-update OS image. But something like that (10k tests on physical machines, millions for VM's) is what I would _expect_ they are doing.

Previous discussion:

>Microsoft suspects some PCs might not boot after Windows 11 January 2026 Update

https://news.ycombinator.com/item?id=46761061

Updated Windows 11 on Parallels: Has been stable since 2021.

Boot loop: Can't system restore. Can't roll back updates. Can't reset PC. Can't even enter safe mode.

All options ran into unexpected errors and cancelled out. Only option left was to shut down or "restart".

Had to clean install and attach the old virtual drive to a new installation to copy files across then copy the new installation's disk back to the old system as a replacement to ensure it was able to activate.

Seriously considering if I even need Windows anymore.

I'm still on Windows 10, and the only issue I have is the huge nag-screen promoting Windows 11 that appears sometimes when I boot up.

No option to say "don't show again". If anyone knows how to permanently disable this intrusive Windows 11 propaganda screen, please share. I tried searching for a solution but the one I found - a registry change, didn't work.

That's terrifying, as I currently have no boot stick. Does someone know a reliable free system backup tool for windows, in simplicity comparable to Timeshift on Linux Mint, which I can start from an USB Medium to restore a broken system? (I need to able to exclude some folders, like Steam games)
LLMS are a farce of productivity, one of the largest companies in the planet with Azure (cost reduction) AND their own inhouse chips and LLMS still can't get anything meaningful out of it, trillions of dollars spent, Lol.
>nightmare gets worse

Gets?

It was actually just as bad when first deployed as it is now, but none of the key humans who were supposed to know about things like this in advance, knew about any of it in advance.

That's the approach that makes it the gift that keeps on giving.

Or the embarrassment that keeps on embarrassing.

Is there a person or team having high standards that is able to accurately say when the changes introduced by this particular download alone have been thoroughly reviewed to their satisfaction?

Or will there ever be anybody like that ever again?

I tried Win11 out for maybe a week before upgrading (I refuse to call it a downgrade) back to Win10. Explorer of all things being slow as molasses on my very beefy system was the straw that broke the camel's back for me, but also idiotic decisions like not letting my move the damn taskbar without hacky 3rd party software certainly didn't help me feeling frustrated with every minute of use.

I dunno what the fuck they're smoking at M$ HQ, it's truly baffling.

Guess 30% AI generated code isn't good enough.
So happy being on supported, stable version. Every other Windows release is a nightmare and when I saw W11 empty task manager (yes, the bug still persists, of course no response from M$ and some private contact saying to "just reinstall it"), bootloop after defender signatures update and other funny stuff - waiting for Windows 12.
i hope that whoever caused all those bugs doesn't move later into developing Azure.

(part of me actually wishes it would happen, ngl).

It didn't need Ai to create something as horrific as windows vista.
Why doesn't Fedora suck, but windows does? Serious question.

Why is windows literally worse than Fedora? I'm not exaggerating, I just can't understand.

As they wrote it. It only affects physical computer. It seems we now have to learn that win11 is only for VM. Lesson learned
no matter the industry, quality control isn't a tool. you can find tools to produce content and to help test for quality, but the ultimate bar for quality is depends on team members.

The issue is that despite code assists (pre and post AI ) helping to produce more testable product, the bar for quality acceptance continues to decline.

Why is windows so hard? In my many years of Linux, I've never managed to brick a computer. Microsoft makes computers hard for no reason. At worst, in the olden days I used to just boot into a livecd and fix my issue, including using an old kernel. Today, I just revert to an old zfs snapshot or if something is truly awful just pull my archived zfs snapshot.

I mean obviously windows can be reinstalled and restored, but my nixos desktop flake can be restored in like 10 minutes while a windows install takes hours

It's 2025... Why are we still dealing with these problems?

Look on the bright side; at least "not booting" is better than "deleting all your files": https://news.ycombinator.com/item?id=18189139

Also, every time MS fucks up an update, more users will become persuaded to turn them off completely. It's a massive amount of trust and valuable user time lost. They keep harping about how much cyberattacks cost, but are clearly silent on the cost of periodically breaking everyone's PCs in various ways.

Interesting. I bought a brand new Windows Arm machine the other day that was DOA: It booted with the UNMOUNTABLE_BOOT_VOLUME message. I brought it back to Microcenter and exchanged it and the replacement works fine. I wonder if that machine was just updated at the factory before it shipped...

I'm a longtime Microsoft fanboy, but even I wait a couple of weeks before updating anything, unless there's an actual problem I need the fix for.

Vibe-coding at scale.
I can't wait for vibe-coded TPM implementations.
> It's unclear why January's security update for Windows 11 has been so disastrous. Whatever the reason, Microsoft needs to step back and reevaluate how it developers Windows, as the current quality bar might be at the lowest it's ever been.

I think I might know...

They just keep digging