Microsoft CEO says up to 30% of the company's code was written by AI | TechCrunch
Microsoft CEO says up to 30% of the company's code was written by AI | TechCrunch
Now I understand why at each windows 11 update, they introduce more bugs than ever
This is my own experience but the past few years Windows has been extremely dependable for me and then in the last few months the updates they’ve have been terrible. I’ve seen more blue screens recently than I have in a lot of years.
All this to say that if it is 30% AI code being used then it’s very telling!
Government spyware finally has a challenger for the title of "primary reason that most Microsoft software runs like hot garbage".
So the CEO is trying to tell investors that they are saving money by not paying employees. But to me it sounds more like: we are letting our sub-par products continue to enshitify, and any other company using AI to program will be equal competition.
I think he's trying to say that their AI writes code good enough for Microsoft. Which is a message to other business leaders that your company too can benefit from copilot, just hand over your credit card!
Microsoft has absolutely gotten worse in the consumer space, but that isn't really their business these days.
Good point
What products do they have that are enshittifying?
Here is the definition: https://en.m.wikipedia.org/wiki/Enshittification
Have you not noticed that the Windows search has become a meme for being really useful in windows seven and useless in newer versions because it started
And that's just one example that's obvious enough to become a meme
BTW: any form of making a Microsoft product worse for profit of Microsoft is enshittification since they have both endusers and sellers of products that only work on windows/in the Microsoft ecosystem locked in with significant costs tied to leaving.
Here's the list of what they are enshittifying: https://en.wikipedia.org/wiki/List_of_Microsoft_software
Windows 11 was much much shittier, idk about the rest of their products, I haven't used them.
And I switched over to steamOS, so I won't be using any of their products again.
What products do they have that aren't?
this makes way more sense than hundreds of shitty devs.
Even their AI crashes all the time, its brutal.
this is why I get so much business as a IT consultant lol
We know Microsoft.
Hardly possible not to noticce..
“Written by software” does not inherently mean AI.
Code generators being a prime and ancient example
I vaguely remember talk about companies like Google having software that fixes/writes code and that was ages before LLMs.
Well that explains a lot
Boy am I glad not having to touch their software.
Year of the Linux desktop
(Amusing sidenote: my autocomplete's first suggestion after 'Linux' was 'propaganda'.)
Sure as hell feels like it!
Everybody saying this is why their products are shit are really confusing me. It's not like Microsoft just started being terrible. They've been terrible for a real long time. Way before AI was a thing. This is just a symptom of Microsoft's awfulness not a reason for it.
Ok, it's like this.
Ms used to release shitty stuff. And they'll continue to release shitty stuff except now it'll be 30% more shitty.
The AI is not the reason their software is bad, but their software is bad for the same reason they’re claiming to use AI to write it.
There were alpha versions of windows 8 with less glaring/annoying bugs than windows 11, though
My windows 11 gaming machine has done all manner of fucky stuff, including permanently losing desktop icons seemingly at random and just whole ass refusing to open the file explorer for six months.
Windows 11 is so terrible so far that if I'll need to use Windows 10 for dev reasons, I'll either pirate the extended support patches, or use a shitbox (obsolete PC for optimization purposes) disconnected from the internet. I do fear that I might have to hack a GUI onto LDB or GDB, because I got too used to RemedyBG (I'm already using Kate).
And 100% of their product names are generated by AI. No human would be stupid enough to change something's name for zero reason.
No human would be stupid enough to change something’s name for zero reason.
Well, unless you're a product manager at Google apparently... Though with them you're lucky if its just a change of name rather than it becoming an entirely "new" thing, or just getting outright axed...
coughs
That's why their products are so crappy!
What? There products have long been shit ever before AI was even a thing.
Anyone remember windows ME? I sure as fuck do.
No way, they get their results through honest effort. Anyone can make crappy AI products now, but Microsoft have been doing crappy products the hard way for decades. Don't downplay their hard work !
50% of my code is written by Intellisense...
If they start with those products today with zero marketing budged and zero user base nobody would use it. Those CEOs are just clowns.
Even worse, whenever a good new technology does pop up they buy it and ruin it.
Windows hate train looks fun, but as someone who works in the industry, most of that code is probably just unit tests and boilerplate stuff.
Copilot is decent at quickly writing huge amounts of mostly correct, tedious unit test code, depending on your language/framework. And since Microsoft works with languages like C# and .NET for their native apps, and likely backend too, there is quite a bit of verbosity that Copilot can take care of. Also, documentation might count as well.
No real code is AI-generated. He's just saying shit like this to keep idiot investors happy.
Thanks for the context
Isn’t that the problem tho. He’s the CEO of Microsoft which is supposed to be a bight end technology firm saying bullshit
Operations are one thing, but investors are another. The latter don't know the sector, but they want profits.
So you have to convince them not to interfere with your activities, because they can make things worse.
Yikes
It shows
Beat me to it.
That explains so much
Code written by software doesn't mean AI unless you ignore compilers
Executives lie to boost profits and justify their decisions, I doubt MS execs even know how much of their code is AI generated just like the ad sales company exec they were talking to in the article
This makes sense and would explain the mainline windows versioning and probably the xbox versioning too!
Microsoft to AI: List all the integers from one to eleven.
AI: 3. 95. 98. 2000. XP. Vista. 7. 8. 10. 11.
Well, that explains Windows 11.
They're attempting to make excuses for their inability to create functional software
They’re lying about using AI to write software, they probably have required all their programmers to have an AI plugin installed, and are thus counting any code they make as “written by AI”, and then are counting any minor edit to existing code as the entire thing being “written by AI”.
The software is bad because it’s written to serve the infinite growth imperative. The reason they claim they’re writing code with AI is because that being true is the only hope that they have for achieving the infinite growth imperative. It’s a con, it’s a cult, they are extracting as much value as they can before everything falls apart.
Not suprised
That’s… not something to be proud of.
He used the words "written by software". This is ambiguous and doesn't mean AI, for example, using annotations for variables and generating the getters and setters would count. Right click and create function body for interface function definitions also.
They're exaggerating to pretend their AI is more useful than it is.
Intellisense in visual studio has also been really good for over a decade. Which is technically also written by software and not me.
I mean, really good intellisense is a great improvement, but it's not replacing devs any time soon.
People have been using annotations to generate code since I rode my dinosaur to work.
Don't forget code generation for stuff like bindings or database schemes
I'm still forced to use Microsoft Outlook and teams, unfortunately, and boy oh boy is it bad.
Yesterday i spent 45 minutes of a 1,5 meeting (that would have been 45 minutes) on trying teams to please try and use the right microphone, please share a screen (not working under Firefox or chrome now, apparently)
I can't wait for the day that I have some time to get us off that dog shit
I love that Outlook occasionally fires up one of its keyboard shortcuts and clears your entire email you were typing if you're not paying attention.
Fucking love it.
not working under Firefox
Weird
or chrome
Oh, there it is
Feature exclusive, huh!
Outlook is pretty damned bullet proof, Teams, OTOH, is a fucking mess. I can see IT wanting to keep everything in the same ecosystem, that's perfectly sane, but I'm certain Zoom can be setup to honor AD credentials. We set it to use Google SSO.
Ha, yeah my Outlook will randomly just not be able to connect to server when starting up. This is from fresh boot every day, some days it just can't, have to reboot and then it works! Fucking brilliant, they managed to break email.
Horseshit.
The current state of code generated by AI is sketchy at best. I often get plain wrong answers because the model tries to derive. It comes up with calls to functions and properties that just do not exist.
"You are right, I made a mistake. Here is a better answer." Continues to give wrong answers.
Apart from that, apps that are glued together from AI generated code are not maintainable at all. What if there is a bug somewhere and you so not comprehend what is actually happening? Ask AI to fix it? Yeah good luck with that.
I do use AI for simple questions, and it works fairly well for that, but this claim by MS is just marketing bullshit.
"You are right, I made a mistake. Here is a better answer." Continues to give wrong answers.
To be fair, the AI's not wrong. It's probably better, but just a teeny tiny bit so.
Honestly, AI is like a genie - whatever you come up with he'll just butcher and misinterpret so you start questioning both your own sanity and the semantics of language. Good thing these genies have no wish limit, but bad thing that they murder rainforests while generating their non-sequitur replies.
This ^
"20%-30% of code inside the company’s repositories"
Now, if they had said "20%-30% of code written in the past 6 months..." I might buy that.
The repositories are going to have all the current codebase, likely going back years now. AI generated code is barely viable at this point and really only pretty recently.
No way 1/3rd of all current codebase is AI.
Even 20% of new code would be a stretch unless they count every first iteration of code written by AI that needs to be replaced by a human later because it was plain wrong.
"Please move all comments from in-line to the line above, and add a separator line"
They mean new code, as per the article. And they mean code gen and IntelliSence
"Auto complete generated 30% of characters"
Fixed it.
"You are right, I made a mistake. Here is a better answer." Continues to give wrong answers
The exact same wrong answer. Co-Pilot is especially bad for that. I'm practically giving up using it outside of vs code because the actual copilot AI is dog shit stupid m
I didn’t RTA, but if they mean ALL code at MS, that just can’t be true. They have legacy stuff going back decades, beyond just their windows platform. There’s no way 30% of all their code is replaced or newly created by AI.
They say that because they are selling it.
And yeah, my experience is the same. The most frustrating is when writing in a typed python, and it gives answers that are clearly incorrect, making up attributes that don't even exist etc.
My brother said his superior asked him to use more AI auto complete so that they can brag to investors that X percent of the company’s code is written by AI. This told me everything about the current state of this bullshit.
It would explain the constant barrage of system breaking bugs and RCE vulns
Turns out they have been using AI for decades.
Ah but it was nice for a minute imagining microsoft vibe coding windows...
This only makes sense if they are counting intellisense auto complete as "AI written"
Has to be something like that. Nadella is somehow cheating with the number, trying to keep the AI hype going.
You could say ALL of my latest scripts were written with AI. Because I often use it to get a hint or gather some boilerplate code (which I still go over and modify).
Was the auto complete in visual studio not a "trained" set before the llm craze kicked off? Would not surprise me if they decided to include that.
They include tab complete of github copilot which is often as much as a single dot. Same thing they've done with all github copilot stats.
If this were true there would massive databreaches. AI is really bad at keeping private keys private. Not to even mention the default credentials it would use because it doesnt have commen sense to change them
Likely a lot of manpower were focused on that, and/or the employees rather wrote their own code then lied about the AI use (heard a lot about it).
I disagree. It feels like your making this assumption from the point of view that people using AI to develop turn their whole brain off and let AI take the wheel. Any dev I know using AI uses it as a time saving measure, i.e. advanced autocomplete, or to assist with troubleshooting as a form of advanced search engine. Also you would have no need to give the AI the actual key itself, at most you would give it the title of the variable the key is saved as.
I find it hard to believe because I work at an adjacent company who has made similar claims and it is complete bullshit.
I do think there is some about of "AI provided a smart complete and the developer hit 'tab' to take the changes" equalling "this code was written by AI" in some metrics that go to the execs. And since the execs mandated high AI use everyone is fine just saying they have high AI use regardless of how true it is.
Yeah that'd explain some stuff. Happy to have switched to Linux
Same. My games even run faster.
Fun fact: Nadella has been replaced with an AI agent a couple of months ago and nobody has noticed yet. "Copilot, while I'm away, generate bs on AI adoption and fire a bunch of employees, ok?"
So this explains why Microsoft Swiftkey is total dogshit now. Also why the Outlook app barely works.
Its unbelievable.
I used to be able to swipe freely on SwiftKey, and now I can't really do it without being extra careful and mindful of not spelling the wrong thing. Idk what Microsoft did to the product but I wouldn't call it an improvement.
Idk what Microsoft did to the product but I wouldn't call it an improvement.
I think the article we're looking at here isn't really hyperbolic. They got AI to write all their code and broke the Keyboard.
Just FYI, if you can live without swipping, I recommend FUTO keyboard., it is basically Swiftkey but it actually works and doesn't come with Microsoft's spyware built in.
It's what I use now, and I'm really happy. Don't be fooled by it being in Alpha, because it works flawlessly (minus the swipe, which is hit and miss).
Eww. Maybe it’s not really true and Microsoft just wants to remind us that big corporate AI is so legit that all the software you use all day was “helped” by it.
But really for me the issue is the company, not the AI. If I read an article about AI generated code making it into the Linux kernel or some gnu/kde/etc utilities, I don’t think I would worry much because those changes will be reviewed by cranky old nerds who care about the functionality of the software first. I have no such confidence in Microsoft’s processes.
Power move by the zucc by first asking how much genai is used at Microsoft then refusing to answer his own question at Facebook 😂
And how many times was all that code rewritten?
I bet they’re counting code written while someone had an AI plugin installed as “written by AI” and I bet that accounts for almost all of that 30%. On top of that, I’m betting that they made it mandatory to have such a plug in, and the other 70% is just code written before they mandated this.
I would be very surprised if 30% of their code lines had even been touched at all by anyone since AI coding assistants became a thing.
I wish this shot from The Terminator had the camera showing Sarah Conner's face instead of Reese's, because it'd be such an appropriate meme image on multiple levels for when someone makes a misleading claim about some current AI system.
I could see stuff getting small changes and them claiming that the entirety of the new version is “written by AI”.
Also, having 1/3 lines with obvious code that can be auto suggested correctly would make sense, but that is hardly code "written by ai" in the way they suggest.
I’d guess a lot of the people writing the code don’t even have it turned on, it’s just installed because management said it had to be, because management wants to be able to tell investors they’re “innovating work flows”.
Those are the easy time savings though, the safe easy stuff the developer doesn't have to worry about anymore. (Giving them time do the gnarly stuff)
Well, that would explain a lot.
I'm also guessing that at "up to 30%" of the company's leadership decisions are being made by AI too.
Honestly, I'm not surprised.
??? No it’s not! Can investors sue because this is such an obvious lie? Pls I have 0.3 Microsoft shares
Are they including code generated to test their own models capability maybe?
Would be interesting to see how they measured that metric. Are they tagging individual lines as AI generated?
What those lines are too would be interesting, AI as auto complete is less dangerous than complete generation, but probably also less useful.
Most probably Microsoft has set objectives for how much LoC are from LLMs and developers invented numbers to match that metric (because they probably have things more important to do than counting LoC)
AI as auto complete is exactly what I was thinking.
I've seen lots of cases where AI appears as an auto complete suggestion and I can just hit
<TAB>
and it finishes the current line. It's essentially filling in the boilerplate text. Heck in some cases it isn't even right, but it's close enough that I can change a few values.I also want to point out that this isn't particularly new technology. This existed before AI. It has perhaps expanded more, but it isn't a revolutionary improvement, it's an incremental one. So when we talk about usefulness, I think it is actually more useful.
Now if it could do all the magic planning and thinking, that would be more useful, but we're not there yet.
I hope that that's inclusive of something like lines of documentation in comment lines.
