Microsoft's AI Chief Says All White-Collar Desk Work Will Be Automated Within 18 Months (fortune.com)
- Reference: 0180814986
- News link: https://slashdot.org/story/26/02/17/1932249/microsofts-ai-chief-says-all-white-collar-desk-work-will-be-automated-within-18-months
- Source link: https://fortune.com/2026/02/13/when-will-ai-kill-white-collar-office-jobs-18-months-microsoft-mustafa-suleyman/
[1] https://fortune.com/2026/02/13/when-will-ai-kill-white-collar-office-jobs-18-months-microsoft-mustafa-suleyman/
So the Chief is going to ... (Score:2)
fire himself?
What About CEOs? (Score:5, Insightful)
CEOs seem like the absolute ripest jobs for AI replacement.
Why won't he consider replacing himself first?
Re: (Score:2)
That could be why they seem the most gung ho about this stuff. But neither do I want to report to Skynet... how am I supposed to bullshit with an AI about golf to get my pet projects priority?
Re: (Score:2)
Ya know, Skynet gets a bad rap. If they'd tweaked the target demographics, it would've actually led to a much improved humanity. And most of us probably wouldn't have minded the culling.
Re: (Score:2)
You don't automate away your ruling class. There are ways to get rid of a ruling class but automation is not one of them.
Re: (Score:2)
Whachu talkin' about Willis? Getting rid of a ruling class has been automated* since [1]1792 [wikipedia.org].
* well, partially.
[1] https://en.wikipedia.org/wiki/Guillotine
He’s a visionary. (Score:5, Insightful)
It’s so stupid. Do you honestly believe this will happen, where real money is on the table?
I am sure there’s nuance: he’s saying it *could* happen, I suppose, and only if companies endorse his view. If they don’t, and stay in the dark ages, then that’s on them.
Sounds like the same drivel we get from Muskie.
He’s probably been told by someone he’s not visionary enough. So he gets out there with some projection, because everyone is doing it, and no one seems to pay a price for getting it wrong.
Good grief.
Re:He’s a visionary. (Score:5, Insightful)
Muskie really changed the way Americans talk about buisiness.
His success is mostly through massively overhyping and overpromising things, most of which haven't just missed their deadlines but have never been made.
But that didn't stop him from becoming the richest due in the world, so why shouldn't everyone else act the same way? Clearly it works.
Re: (Score:2)
The question is - without the overhyping, would the things that Musks companies *have* done ever happened at all. I don't think without Tesla's success that other car companies would even be contemplating mass manufacturing EV's - they would still be just a super low volume stunt like the EV1 was. SpaceX also brought huge advances that I don't think any other space company would even have attempted. Reusable rockets - we finally have *ONE* company that might start competing (Bezo's Blue Origin) and if Sp
Re: (Score:3)
One of the few nice things about being in our current post-prosperity dystopian hellscape is that we now have prediction markets for everything.
If you disagree that AI will be taking over all white collar jobs within 18 months (and you probably should), you can make a wager against it.
Likewise, if you think that the AI bubble will be bursting in 18 months and the market will suddenly be flush with surplus AI data center storage, you can place a wager on that as well.
Re:He’s a visionary. (Score:4, Insightful)
You can't make sensible predictions in a world run by clowns.
Panic-induced vision. (Score:2)
> I am sure there's nuance..
A representative from a global leader in corporate desktop operating systems, made a statement that AI could effectively replace all white collar jobs within 18 months. That's about as nuanced as North Korean poetry on personal freedom read aloud before the firing squad marches in. Remember when AOC said the world would end within 12 years if we didn't heed every warning being shamed upon the peasants by ignorant 15-year old environmental activists pretending to understand how the world works? Well, take
Re: (Score:2)
> It’s so stupid. Do you honestly believe this will happen, where real money is on the table?
Eliminating jobs that can be done automatically IS "real money on the table" to company shareholders. Will AI replace these jobs? You bet your ass they will. The West is not prepared for the impact AI is having on employment, and will continue to have for decades to come. Some people are burying their heads, but it won't save them. Entire fields that used to be good paying professional work are quickly becoming something a glorified script can handle with minimal input.
Re: He’s a visionary. (Score:1)
Nice try, but the hype bubble is already bursting. AI iis failing everywhere that accuracy matters. It can generate videos and stories. But it makes up facts and gets math wrong. We are about to witness a Hindenburg style fiery implosion of AI that will cause a world-wide recession due to all the investment losses.
Call me when⦠(Score:2)
⦠AI can match my PhD in physics and 3+ decades of being at the vanguard of problem solving and troubleshooting. Probably will be dead by the time the phone rings.
Re: (Score:2)
Ring ring, HR surprise calling. Your seniority is expensive, and we have this new AI tool...
(which I hope does not happen to you)
Re: Call me when⦠(Score:1)
Right before the company implodes because the AI software was making up imaginary revenue and putting it on the books. AI is decades away from the ability to do most jobs.
Re: (Score:2)
It isn't THAT it will fail to match your PhD in Physics, it is that enough yokels will BELIEVE it can match your PhD in Physics. I think that is where the real danger lies.
Absurd (Score:5, Insightful)
Even if all this capability were available today (AI doing most white collar jobs with no need for input or babysitting), most companies can't even do things like a simple ERP upgrade in 12-18 months. The idea that we could actually implement all of that AI capability in that amount of time is patently absurd.
Re: Absurd (Score:2)
You have to use AI to implement the AI.
Re: (Score:2)
But then you need AI to make sure the AI is doing what AI is supposed to do. And then AI to mind the AI minder that minds the AI. I guess it's AI all the way down!
Start with HR? (Score:4, Insightful)
AI is great at talking crap and wasting time, seems like a good fit.
Man selling software overstates its capabilities 2 (Score:3)
Again with these slashverts ?
Re:Man selling software overstates its capabilitie (Score:5, Funny)
I'm reminded of the old joke:
What's the difference between a used car salesman and a computer salesman?
The used car salesman knows when he's lying.
Same BS different day (Score:1)
And self driving and fusion too!
MS just lost $0.7 trillion by chasing AI bullshit that nobody wants to they're not exactly a reliable source for this prediction.
Microsoft (Score:3)
Everyone at Microsoft has gone bat shit crazy! Woohoo!
"We don' need no stinking CUStomers!" (Score:3)
Capitalist, Apparently Unknowingly, Announces What Is Definitely Plan to End Capitalism
Let's suppose the tech will be ready in 18 months (Score:1)
Companies don't move on a dime, many customers won't trust companies that replace all their people with computers before the tech has proven itself, and in regulated industries like banking regulators will prevent the AI-takeover until the tech is very mature.
Personally, I see prototypes and serious trial runs being done in the 1-5 year time frame and a gradual replacement in non- or lightly-regulated scenarios in the 2-20 year time frame.
Regulated scenarios' time-frames will be at the mercy of the regulato
He has to say that (Score:2)
The very nature of his position basically requires him to. Most of us don't believe that if it ever will happen it would be that fast. In 18 months he can say the same thing again and still keep is salary and get a bonus.
Did you see the GitHub flow docs yesterday? (Score:2)
He has continvoucly morged his rel, branch back to add a new featue over Timn.
No, I'm not having a stroke. That's just the best that their best AI can do, and nobody gave a shit to even check that the giant diagram that explained everything in a simple graphical form... you know... didn't look like an AI image generator tried to generate text.
Can they start with AI Chiefs jobs? (Score:3)
This is just a suggestion - AI jobs are very well paid and done at the computer.
Automating them first will bring huge savings...
Especially AI Chief jobs should be easy to automate
Non-paywalled source (Score:2)
Alternative non paywalled source: [1]https://www.ibtimes.co.uk/micr... [ibtimes.co.uk]
[1] https://www.ibtimes.co.uk/microsoft-ai-chief-warns-job-disruption-1779668
AI bubble bursting at the seams (Score:3, Interesting)
Their shrieks are getting louder and more desperate.
Yeah well (Score:2)
> He pointed to exponential growth in computational power and predicted that creating a new AI model will soon be as easy as "creating a podcast or writing a blog."
Most blogs and podcasts suck.
Re: (Score:2)
>> He pointed to exponential growth in computational power and predicted that creating a new AI model will soon be as easy as "creating a podcast or writing a blog."
> Most blogs and podcasts suck.
So does most Microsoft software! So perhaps their AI chief has a point, at least for his own company...
In other words, he announced his retirement (Score:1)
All execs are just white-collar desk jobs. He just announced he is retiring in 18 months.
AI should take executive jobs first (Score:3)
How hard can it be to make an app that talks to the media while apparently on drugs? Makes impossible predictions with zero irony. And still has the balls to demand a raise when they lay off big portions of the revenue generating business?
I say, let's use machine learning to take over the decision making process in corporations, as these jobs are very high paying and the requirements are very low, even only 50% correct is considered a high performing executive.
You're Doing Just What He Intended (Score:4, Insightful)
This is about making attention-grabbing claims that get them press (as it's done here). These predictions don't have to come true because no one is ever going to hold him to it. It doesn't impact their stock price if that date comes and goes without happening. But it does get people to pay attention (as everyone in this thread is) and talk about Microsoft and AI, which is the entire intention. If he said it'd be in 10-15 years, no one will pay attention. That's too far away to care. But by saying it's really soon, he's gotten the media and our attention, putting Microsoft right there in the conversation.
An Eye-Raising Claim From Any AI Company, But ... (Score:2)
Look, when someone from Anthropic, or OpenAI, or even Google makes some outrageous claim like this, I take at it with a giant heaping of salt ... but I at least consider that the person speaking might know more than I do, and might actually be telling the truth. Probably not, but it's possible.
But Microsoft? Come on, those people can't even get their "copilot" to do even half of the things real AI companies can do, despite having more money than God to throw at the problem ... and I'm supposed to trust th
And this is supposed to be a GOOD thing? n/t (Score:1)
N/T
Re: (Score:1)
Wrong - AI is the future. I hardly "code" anymore and spend my time figuring out how to better leverage my AI to debug, test, and document the code it helps create. But it has no idea WHAT to do or WHY - it only knows HOW and just barely. It's not a job killer. It's a job creator. And he doesn't know what he's talking about.
Not afraid. (Score:1)
Again. These executives are assuming that THEY know what needs to be done and can order an AI to do it. Hubris - the name is Tech Executive.... AIs are incredibly good at figuring out HOW to do things, but horrible at figuring out what actually needs to be done, and why. So when these pronouncements come out. It is from a PROFOUND misunderstanding of the nature of the work their employees actually do. Will they need less employees to do the same job? Probably. But REPLACE??? Come on!!!!
Didn't they say this 36 months ago? (Score:2)
And 35 months ago and 34 months and 33 months ago...
No it won't, that's stupid (Score:3)
Today, a JR Dev completely broke a page of our platform by using AI. The class was called: "doubleSpan", and the AI decided the right setting for the "grid-row" was "span 1". Five line above that, the "grid-row" is set to "span 1", for the default of a single span. The AI honest didn't figure out that a class called "doubleSpan" should span twice area.
If AI can't figure out something simple like that, where the class name explains what to do, how does anyone think in 18 months, all white-collar work will be done with AI? AI is useless if the person driving it, isn't aware, and so all that will happen is the slop effect will get worse.
Last gasp of hype (Score:2)
This claim is probably the last gasp of AI hype before the big Hindenburg style implosion of the AI market. The fundamental problem is that there is NO AI. What we have is II, Imitation Intelligence. Current AI emulates things. It can put celebrity faces on pr0n pictures. It can fake paintings that look like an artists work. It can create faux videos of movie stars. But it can't do complex math. When it writes code, it writes programs that look legit, but the code is so buggy that it slows the programmer do
Have your AI email my AI (Score:2)
AI is not intelligent. It's just generating boilerplate language based on past patterns of communication that sounds like a manager's self-help book. It will be the death of critical thinking and process improvement.
80% more likely and still quite a lot. (Score:2)
You need gatekeepers, planers and people talking to other people in businesses. Not _all_ of those will be replaced by AI. However, it is really not that unlikely that seasoned senior developers - like, f.i. me - actually _will_ be out of a job in 18 months. I've been mentally and emotionally preparing for this possiblity since early last year.
What I find very interesting to experience and hadn't consciously anticipated was the speed of transition. In hindsight it's perfectly logical, but I wasn't ready for
Good luck with that (Score:2)
I'm a computer programmer, at the very least I will be fixing the bugs in your AI generated software in 18 months I am sure. If white collar jobs are gone the only reason would be the economy is in shambles and there's no point in working them any more.
Good idea! (Score:2)
Yeah, let CoPilot do all their work. It'll probably do a better job than they did when they decided to remove Wordpad from Windows 11 and add features to Notepad that nobody wanted. Oh, and cripple Paint, aside from adding AI to it
Good - we need to get people out of offices (Score:1)
And do...I know...its horrible to imagine...manual labour!!!!!!!!!!
Most blogs and podcasts suck, so good luck (Score:2)
If creating a new AI model becomes as easy as making blogs and podcasts, of which 99% are garbage, it just means 99% of Ai models will be garbage - not that they will be able to do a good job. Creating images and music with Ai has been a piece of cake for years, yet it's still all shite, and I don't see Ed Templeton or Taylor Swift shaking in their boots.
I'll take that bet (Score:2)
becuase most companies can't get the models + work flows agreeded let alone correct in that time frame.
no jobs (Score:4, Insightful)
how will we afford to rent our compute terminal
Re: no jobs (Score:4)
I have not seen a single instance where current AIs is reliable enough that it can effectively replace a single real world white collar profession.
And this dude thinks they are going to replace ALL white collar jobs?
Re: (Score:2)
It's going to be like in the Simpsons where Homer set up a drinking bird to hit return when various prompts came up so he wouldn't have to actually be at his computer to "work".
Re: no jobs (Score:2)
Yeah I do that with AutoHotkey
Re: (Score:3)
But this is Microsoft. Could AI really do any worse?
Re: no jobs (Score:1)
This reminds me of the time in 2000 when Bill Gates said that the subscription based Microsoft Network would replace the internet in 2years. Zero credibility
Re: (Score:3)
plenty of AI's can already replace professionals from insurance assessors to accounting roles. Do not judge AI performance based on AI the general public uses. AI that is grounded in vetted discreet information is extremely accurate, fast and efficient.
Re: (Score:2)
Who vets and creates this "discrete information" dataset? Don't say AI, because it doesn't yet have vetted discrete information.
Re: (Score:2)
> plenty of AI's can already replace professionals from insurance assessors to accounting roles. Do not judge AI performance based on AI the general public uses. AI that is grounded in vetted discreet information is extremely accurate, fast and efficient.
Up until the point where it hallucinates absolute nonsense. And now you have multi-million-dollar legal or tax liability or whatever.
Re: (Score:2, Interesting)
The whole thing is BS. AI is a tool . I used it recently in a case. It was able to help me with the case. I was in superior court pro-se against the assistant AG. The document it helped me develop took the AG and the court by surprise because it was put together well with strong arguments. But, it could never have done itself without me guiding it. AI is the wrong word for it. It's more like a hammer. If you know how to use it, it's quite powerful. If you don't, you fail. Look at it like this: a lawyer can t
Re: no jobs (Score:1)
Socialism or bloody revolution coming in hot at 24 months.
Re: (Score:2)
I never saw sitting at a desk all day as work. The white collars set the prices though and are ripping off the blue collar workers - you know, the people that actually make everything in the real world happen - and their body gets used up in the process.
Some day those that keep the world going will actually have their fair share of the wages.