I was, unfortunately, one of the engineers impacted by this. From my perspective, it was as poorly handled as this article paints it. I got a meeting invite at noon for a meeting at 12:30, which I...
I was, unfortunately, one of the engineers impacted by this. From my perspective, it was as poorly handled as this article paints it. I got a meeting invite at noon for a meeting at 12:30, which I didn't even see because I was out for lunch.
Turns out a bunch of my team got cut with no announcement or acknowledgement, just a few short days after we got an email noting how well our product performed during the past quarter.
We were already stretched so thin on resources that product quality was negatively impacted, so I have to imagine this is really going to hurt them over the medium-term future. AI is nowhere close to replacing engineers - I'm not sure if senior leaders really believe this or they're just using it as an excuse to squeeze.
Ouch! Hope you land on your feet! I don't think AI is really to blame. This seems more like cost cutting. From a business perspective: You have a bunch of talented devs and (presumably) a long...
Ouch! Hope you land on your feet! I don't think AI is really to blame. This seems more like cost cutting.
From a business perspective: You have a bunch of talented devs and (presumably) a long backlog of features to build and bugs to fix. All of a sudden you are able to increase the productivity of all those devs via AI. Then logically you'd produce more features and fixes.
So assuming that Microsoft isn't done with their products this seem more of a cost cutting measure, not an investment in AI to replace devs. Sure they MIGHT use AI to offset the loss of productivity due to the cuts (with various levels of success), but if the goal was to improve their products then getting rid of people working on them seem extremely shortsighted. So to me this is just regular cost-cutting...
Thanks! The one thing I can tell you for a fact is that, at best, AI has only marginally increased productivity for the trams I was a part of. Though I honestly doubt it's improved at all when you...
Thanks!
The one thing I can tell you for a fact is that, at best, AI has only marginally increased productivity for the trams I was a part of. Though I honestly doubt it's improved at all when you factor in all the time spent fiddling with it.
sigh good luck dealing from the fallout of this. I work with a lot of legacy code from someone who absolutely should never have touched any programming language. I can only imagine how much pain...
sigh good luck dealing from the fallout of this. I work with a lot of legacy code from someone who absolutely should never have touched any programming language. I can only imagine how much pain it's going to bring when this is happening in Microsoft of all places.
For one thing, cutting of so many software engineers when we're still learning on how to use AI as a society properly seems really risky to me. But I'm also surprised at just how small the amount...
For one thing, cutting of so many software engineers when we're still learning on how to use AI as a society properly seems really risky to me. But I'm also surprised at just how small the amount of middle managers being cut is. It's like the top is expecting more for less which rarely goes well. And if anything laying off the ones who're actually producing the product seems, to put it politely, a suboptimal business strategy with an end goal that may be challenging to reach.
I'm now also glad to be a software engineer at a small company... If this trend keeps up, man that's brutal. Stupid, and brutal.
I'm only halfway convinced AI is the actual reason and not just the scapegoat. Layoffs have been on a significant rise in the past year, year and a half, and this gives them a veneer of legitimacy...
I'm only halfway convinced AI is the actual reason and not just the scapegoat.
Layoffs have been on a significant rise in the past year, year and a half, and this gives them a veneer of legitimacy rather than project weakness regarding revenue and costs.
The company that I work for completely got rid of all junior devs to bet on AI, this is without a doubt the dumbest decision they made since they got rid of all QA (now we have outages once a...
The company that I work for completely got rid of all junior devs to bet on AI, this is without a doubt the dumbest decision they made since they got rid of all QA (now we have outages once a month and clients constantly complaining).
They even hired more Agile Coaches (scammer in my opinion) and double down on product owners that are completely useless; I could replace any product owner I had with a to-do list.
I'm not sure if they expect AI to vibe code everything in a tech company.
The difference between a good product owner and a bad product owner is absolutely massive. I have the privilege to work with a few that are very good at their job and they dramatically accelerate...
The difference between a good product owner and a bad product owner is absolutely massive. I have the privilege to work with a few that are very good at their job and they dramatically accelerate the team.
What's the size of your company? Juniors can often be a burden rather than an asset unless your team is structured to educate and elevate them (and retain them!).
I do agree a good product owner can be helpful, but so far it's been a person between me and his superior, to simply forward what we do or pass the tasks down, never adding or completely...
I do agree a good product owner can be helpful, but so far it's been a person between me and his superior, to simply forward what we do or pass the tasks down, never adding or completely understanding what needs to be done.
My company is a bit over 200 people, as someone who was mentoring some of those juniors I can ensure their work was worth it, one of the projects literary had only one junior dev assigned, but they decided to keep only the PO, the project could have been done with the junior dev but in no way with the PO. And of course we need to cover that position now with a senior that will cost more with no real advantage.
100% agree with everything you've said. This seems like a very short-sighted decision based on promises from completely detached middle management that AI is way more powerful than it actually is.
100% agree with everything you've said. This seems like a very short-sighted decision based on promises from completely detached middle management that AI is way more powerful than it actually is.
Two of the affected ones that were part of my team worked on the payment platform, as you can imagine payments are a core part of our product, they were not let go due to discontinuation.
Two of the affected ones that were part of my team worked on the payment platform, as you can imagine payments are a core part of our product, they were not let go due to discontinuation.
I know people at Microsoft who luckily dodged this carnage but are now scared for their jobs or have lost managers out of the blue…
The Redmond-based software maker on Tuesday announced it was letting go of more than 6,000 employees companywide, representing about 3% of its workforce. In Washington state, Microsoft laid off 1,985 workers.
At Microsoft, employees describe an increased emphasis on AI in their daily work. Some said they’ve felt a shift over the past year, with more pressure to incorporate AI into their role rather than view it as a helpful tool. In April, CEO Satya Nadella said in a fireside chat with Meta CEO Mark Zuckerberg that 20% to 30% of the company’s code was written by AI.
[An affected engineer] said she showed up to her office at 9 a.m., received a meeting invite for 11 a.m. and was laid off on a mass call with about 150 other people.
Unlike in 2023, when Microsoft laid off about 10,000 people, employees didn’t receive a companywide announcement about the sweeping layoffs on Tuesday.
I know people at Microsoft who luckily dodged this carnage but are now scared for their jobs or have lost managers out of the blue…
I've been saying for a bit now that I'm not afraid of AI doing my job effectively. I'm afraid of someone giving it to AI anyhow. I work with a product owner who regularly brings me code be created...
I've been saying for a bit now that I'm not afraid of AI doing my job effectively. I'm afraid of someone giving it to AI anyhow.
I work with a product owner who regularly brings me code be created through "vibe coding". It always looks amazing in the outside, but so far the code has always been a lot of disjointed spaghetti with unnecessary complexity, and always always always some flaws. But it doesn't matter if I go into detail or stay light; the minute I explain to him why the results may not be usable or scalable, in can tell that he thinks I'm just making things up to protect my job. Because he can find AI to tell him I'm wrong whenever he wants me to be, it's now a daily struggle to explain to him that (a) I happily use AI when it helps, but (b) I don't use it where it doesn't help, and of not helpful having a product owner repeatedly asking me to justify why I can't just upload their code into a magical bucket and it works.
It's like the top is expecting more for less which rarely goes well.
It depends on how you define "goes well". If the hall is quality software written thoughtfully and with intent, you're right, it won't go well. If the goal is produce mediocre software that people have to use because you have a successful monopoly, then it will go brilliantly.
So cut the workforce down to a the minimum amount capable of producing a minimally acceptable product. The entire consumer base is worse for it, as are the people who could've done a better job for a comparatively small amount of money. But the C suite will have carved back a penny of profit and it will be a raging success to the small number of people positively effected by it.
The only happy alternative is companies like Microsoft paying more than the minimum for a product that is better than the minimum, but ultimately making the same sales as before. They would have no reason to do that other than to make people's lives better, pride in craftsmanship, that kind of thing. Things that are not rewarded by capitalism and thus don't belong in the boardroom.
While i'm not quite as bleak I do think coders have a bad habit of assuming "this time it MUST be done right" when the history of coding has shown anything but. Error handling, rights management,...
While i'm not quite as bleak I do think coders have a bad habit of assuming "this time it MUST be done right" when the history of coding has shown anything but.
Error handling, rights management, and gestures vaguely at just about everything to do with the web is all still not really well handled nor up to best practices.
Rather than make more reliable code we've just made it easier to restart you devices faster. Rather than develop a better web framework we've just piled onto the "we'll finish it eventually" JS. Rather than maximize the productivity of your coders we've outsourced to 20 other countries coders who copy paste from stack overflow.
I see AI as likely much the same, especially compared to the outsourcing. The snake oil salesmen will swoop in over the next few years, swear to the heavens you can replace your team with AI, management will gobble it up, massive layoffs will occur, products will go off a cliff, new companies will be seen as plucky brilliant up and comers for "actually having a fucking team that knows that they're doing", and the cycle will repeat itself.
There's no doubt AI is a powerful tool. It's faster than stack overflow, about as correct, and not nearly as degrading. Still like stack it often requires knowing how to take the answer given and apply it to your situation, and parse out the "wtf is this, you can't even do that anymore" portions of the answers.
I personally only use it for those occasional small syntax questions (recently it was "ugh i normally dump to a csv, but this needs excel, what's the quick way to do that again?) and it mostly does "fine". I would neeeeeeeeeever code something 100% with AI, but the new standard will be entry level engineers cajoling AI code together with more AI code.
It's WYSIWYG DreamWeaver all over again. Looked like a nice site when you stay within its bounds, as soon as a human needs involved in modifying/improving it you end up negating most of the time...
but so far the code has always been a lot of disjointed spaghetti with unnecessary complexity, and always always always some flaws.
It's WYSIWYG DreamWeaver all over again. Looked like a nice site when you stay within its bounds, as soon as a human needs involved in modifying/improving it you end up negating most of the time savings benefit.
I would only caveat that by saying that as a knowledgeable engineer, I do get great results out of AI. But that's because I'm looking at the code it produces and telling it to retractor until it's...
I would only caveat that by saying that as a knowledgeable engineer, I do get great results out of AI. But that's because I'm looking at the code it produces and telling it to retractor until it's nicely organized and looks like something a human would write.
I know the day is coming when I'm told that that is no longer valuable. And that is my biggest value prop to a hiring manager.
I assume this was meant as a reply to me? The one big downside with such a strategy you're asking for people beginning to switch at some point. Which is no doubt why Microsoft is so...
I assume this was meant as a reply to me?
It depends on how you define "goes well". If the hall is quality software written thoughtfully and with intent, you're right, it won't go well. If the goal is produce mediocre software that people have to use because you have a successful monopoly, then it will go brilliantly.
The one big downside with such a strategy you're asking for people beginning to switch at some point. Which is no doubt why Microsoft is so passive-aggresive with integrating its many services. But it has the big risk of collapsing under its own weight eventually. Especially if maintenance becomes more difficult.
Of course, a pretty cynical part of me could also see some board members being aight with it, hollowing the company out, cashing in, and then moving elsewhere...
And yeah. I think that like how a lot of legacy code out there is absolute garbage, we're probably going to see more sloppy software architecture/maintenance as people will apply bandaid solution after bandaid solution with AI.
It was a reply to you. I'm not sure what happened there. I'm still finding my tildes legs. But yeah it agree with everything you said. I didn't think it's a viable long term strategy, because...
It was a reply to you. I'm not sure what happened there. I'm still finding my tildes legs.
But yeah it agree with everything you said. I didn't think it's a viable long term strategy, because monopoly breeds complacency and eventually, as you said, something that eventually falls over. But it does seem like how these things tend to go and I suspect AI will be similar.
Or in five years that work out the kinks and really can do it with AI.
I was, unfortunately, one of the engineers impacted by this. From my perspective, it was as poorly handled as this article paints it. I got a meeting invite at noon for a meeting at 12:30, which I didn't even see because I was out for lunch.
Turns out a bunch of my team got cut with no announcement or acknowledgement, just a few short days after we got an email noting how well our product performed during the past quarter.
We were already stretched so thin on resources that product quality was negatively impacted, so I have to imagine this is really going to hurt them over the medium-term future. AI is nowhere close to replacing engineers - I'm not sure if senior leaders really believe this or they're just using it as an excuse to squeeze.
Ouch! Hope you land on your feet! I don't think AI is really to blame. This seems more like cost cutting.
From a business perspective: You have a bunch of talented devs and (presumably) a long backlog of features to build and bugs to fix. All of a sudden you are able to increase the productivity of all those devs via AI. Then logically you'd produce more features and fixes.
So assuming that Microsoft isn't done with their products this seem more of a cost cutting measure, not an investment in AI to replace devs. Sure they MIGHT use AI to offset the loss of productivity due to the cuts (with various levels of success), but if the goal was to improve their products then getting rid of people working on them seem extremely shortsighted. So to me this is just regular cost-cutting...
Thanks!
The one thing I can tell you for a fact is that, at best, AI has only marginally increased productivity for the trams I was a part of. Though I honestly doubt it's improved at all when you factor in all the time spent fiddling with it.
sigh good luck dealing from the fallout of this. I work with a lot of legacy code from someone who absolutely should never have touched any programming language. I can only imagine how much pain it's going to bring when this is happening in Microsoft of all places.
For one thing, cutting of so many software engineers when we're still learning on how to use AI as a society properly seems really risky to me. But I'm also surprised at just how small the amount of middle managers being cut is. It's like the top is expecting more for less which rarely goes well. And if anything laying off the ones who're actually producing the product seems, to put it politely, a suboptimal business strategy with an end goal that may be challenging to reach.
I'm now also glad to be a software engineer at a small company... If this trend keeps up, man that's brutal. Stupid, and brutal.
I'm only halfway convinced AI is the actual reason and not just the scapegoat.
Layoffs have been on a significant rise in the past year, year and a half, and this gives them a veneer of legitimacy rather than project weakness regarding revenue and costs.
The company that I work for completely got rid of all junior devs to bet on AI, this is without a doubt the dumbest decision they made since they got rid of all QA (now we have outages once a month and clients constantly complaining).
They even hired more Agile Coaches (scammer in my opinion) and double down on product owners that are completely useless; I could replace any product owner I had with a to-do list.
I'm not sure if they expect AI to vibe code everything in a tech company.
The difference between a good product owner and a bad product owner is absolutely massive. I have the privilege to work with a few that are very good at their job and they dramatically accelerate the team.
What's the size of your company? Juniors can often be a burden rather than an asset unless your team is structured to educate and elevate them (and retain them!).
I do agree a good product owner can be helpful, but so far it's been a person between me and his superior, to simply forward what we do or pass the tasks down, never adding or completely understanding what needs to be done.
My company is a bit over 200 people, as someone who was mentoring some of those juniors I can ensure their work was worth it, one of the projects literary had only one junior dev assigned, but they decided to keep only the PO, the project could have been done with the junior dev but in no way with the PO. And of course we need to cover that position now with a senior that will cost more with no real advantage.
100% agree with everything you've said. This seems like a very short-sighted decision based on promises from completely detached middle management that AI is way more powerful than it actually is.
Microsoft has a vested interest in making sure everyone keeps investing in AI, regardless of its real utility.
What were they working on though? Does anyone know? It kinda seems to me like they laid off a whole entire product that they’re discontinuing.
Two of the affected ones that were part of my team worked on the payment platform, as you can imagine payments are a core part of our product, they were not let go due to discontinuation.
I know people at Microsoft who luckily dodged this carnage but are now scared for their jobs or have lost managers out of the blue…
I've been saying for a bit now that I'm not afraid of AI doing my job effectively. I'm afraid of someone giving it to AI anyhow.
I work with a product owner who regularly brings me code be created through "vibe coding". It always looks amazing in the outside, but so far the code has always been a lot of disjointed spaghetti with unnecessary complexity, and always always always some flaws. But it doesn't matter if I go into detail or stay light; the minute I explain to him why the results may not be usable or scalable, in can tell that he thinks I'm just making things up to protect my job. Because he can find AI to tell him I'm wrong whenever he wants me to be, it's now a daily struggle to explain to him that (a) I happily use AI when it helps, but (b) I don't use it where it doesn't help, and of not helpful having a product owner repeatedly asking me to justify why I can't just upload their code into a magical bucket and it works.
It depends on how you define "goes well". If the hall is quality software written thoughtfully and with intent, you're right, it won't go well. If the goal is produce mediocre software that people have to use because you have a successful monopoly, then it will go brilliantly.
So cut the workforce down to a the minimum amount capable of producing a minimally acceptable product. The entire consumer base is worse for it, as are the people who could've done a better job for a comparatively small amount of money. But the C suite will have carved back a penny of profit and it will be a raging success to the small number of people positively effected by it.
The only happy alternative is companies like Microsoft paying more than the minimum for a product that is better than the minimum, but ultimately making the same sales as before. They would have no reason to do that other than to make people's lives better, pride in craftsmanship, that kind of thing. Things that are not rewarded by capitalism and thus don't belong in the boardroom.
While i'm not quite as bleak I do think coders have a bad habit of assuming "this time it MUST be done right" when the history of coding has shown anything but.
Error handling, rights management, and gestures vaguely at just about everything to do with the web is all still not really well handled nor up to best practices.
Rather than make more reliable code we've just made it easier to restart you devices faster. Rather than develop a better web framework we've just piled onto the "we'll finish it eventually" JS. Rather than maximize the productivity of your coders we've outsourced to 20 other countries coders who copy paste from stack overflow.
I see AI as likely much the same, especially compared to the outsourcing. The snake oil salesmen will swoop in over the next few years, swear to the heavens you can replace your team with AI, management will gobble it up, massive layoffs will occur, products will go off a cliff, new companies will be seen as plucky brilliant up and comers for "actually having a fucking team that knows that they're doing", and the cycle will repeat itself.
There's no doubt AI is a powerful tool. It's faster than stack overflow, about as correct, and not nearly as degrading. Still like stack it often requires knowing how to take the answer given and apply it to your situation, and parse out the "wtf is this, you can't even do that anymore" portions of the answers.
I personally only use it for those occasional small syntax questions (recently it was "ugh i normally dump to a csv, but this needs excel, what's the quick way to do that again?) and it mostly does "fine". I would neeeeeeeeeever code something 100% with AI, but the new standard will be entry level engineers cajoling AI code together with more AI code.
It's WYSIWYG DreamWeaver all over again. Looked like a nice site when you stay within its bounds, as soon as a human needs involved in modifying/improving it you end up negating most of the time savings benefit.
I would only caveat that by saying that as a knowledgeable engineer, I do get great results out of AI. But that's because I'm looking at the code it produces and telling it to retractor until it's nicely organized and looks like something a human would write.
I know the day is coming when I'm told that that is no longer valuable. And that is my biggest value prop to a hiring manager.
I assume this was meant as a reply to me?
The one big downside with such a strategy you're asking for people beginning to switch at some point. Which is no doubt why Microsoft is so passive-aggresive with integrating its many services. But it has the big risk of collapsing under its own weight eventually. Especially if maintenance becomes more difficult.
Of course, a pretty cynical part of me could also see some board members being aight with it, hollowing the company out, cashing in, and then moving elsewhere...
And yeah. I think that like how a lot of legacy code out there is absolute garbage, we're probably going to see more sloppy software architecture/maintenance as people will apply bandaid solution after bandaid solution with AI.
It was a reply to you. I'm not sure what happened there. I'm still finding my tildes legs.
But yeah it agree with everything you said. I didn't think it's a viable long term strategy, because monopoly breeds complacency and eventually, as you said, something that eventually falls over. But it does seem like how these things tend to go and I suspect AI will be similar.
Or in five years that work out the kinks and really can do it with AI.
I’m just some guy, but this seems like a good opportunity for people who make products that make microsoft products work like they are supposed to.
Mirror: https://archive.is/cvKtd
Still not unionized? I guess some people will never learn.