Probably should just replace their software engineers too. I don't mean to throw shade, but automotive software is so bad I'd bet my life that Claude could do better.
Nevermind that the update cycle seems to be 6-10 months for changes like "You can now reset your radio presets directly from the radio settings menu", while bugs like temperature control resetting to max cool every start-up never get fixed.
I kept reading about how bad Android Auto was for years but we finally bought a more modern used car and I can't believe they would ship that experience to customers. I had a week where I just had to unpair and re-pair Everytime I got in the car.
I would love to read about why that stuff is the way it is from the engineers, hmm that might be a good spelunking. I really must be missing something that makes it harder than I think it really should be.
US car software, ya, I've never seen such trash in my life. This said I don't really have any complaints about my Hyundai software. It works, doesn't crash, and does what I want it too.
> agent development, model engineering, AI-native workflows -- point directly at where large-enterprise demand is heading.
I don't understand these words. Does "AI-native workflow" mean vibe coding?
I am now seeing a lot of roles asking for "AI-enabled engineers". And I am not sure what that means either. I am sort of afraid to ask because the answer will probably confuse me even more. Maybe it's my understanding of what LLMs are and how they work that makes these words mean very little to me.
Interesting. I saw contractor rates dropping here too. The ones I saw recently were at pre-pandemic levels so from 6 years ago. The large increase in cost of living from that time makes it even worse. The funny part is that the skills and experience that are being asked for are at a senior level. So I guess that would be senior level vibe coder at junior rates?
Management literally doesn't realize that you can't exactly vibe code a serious project? Maybe they just don't care and it's an experiment. Super low risk I guess, depending on the city they might pay the floor cleaner more than that.
FWIW, I interpreted the article as saying they're not looking for vibe coding, but AI model development per se:
"...In practical terms, GM is looking for people who know how to build with AI from the ground up — designing the systems, training the models, and engineering the pipelines — not just use AI as a productivity tool."
I mean agentic coding is a thing but anyone could learn that in a day or week. So the idea of throwing away people that could be the most productive with ai of course make no sense. But it's big corporations, not everything has to make sense. It's most likely a dressed up pure cost saving framed in 2026-lingo.
I was going to say that sounds like short term gain for long term pain. But I'm guessing if there are any issues in the future, the government would just bail them out.
It's really annoying how at some ASIL levels you need 100% code coverage of unit tests. With AI, all you have to do is to get your agent to generate the tests! Likewise with all the MISRA C requirements. Need your cyclomatic complexity to be less than 10? It's just one prompt away! Now your spaghetti code can easily satisfy the safety requirements with much less effort.
As someone who's worked on this kind of stuff at GM, I don't really get the exuberance in this particular space (not just the comment I'm responding to).
If you want 100% coverage, you just autogenerate the test cases. LLMs can't properly check MISRA requirements, so they're really just a layer on top existing automated checkers. Same for complexity metrics, it doesn't get merged if it violates that (or it's a vendor dependency you won't touch anyway).
If you care about the spirit of the rules, they're not that big a difference. If you don't care, there are already ways to do this. In either case they're an incremental change, not what I'd call a godsend.
I hear this all the time. Why does it matter? Punishing a human for making a mistake does not prevent mistakes, nor does it undo the harm of the mistake. A human saying "my bad, I messed up" and an AI saying "my bad, I messed up" are equally worthless, in a functional sense.
"Punishing a human for making a mistake does not prevent mistakes" This statement suggests you don't believe in some combination of neuroplasticity as a concept or the arrow of time.
Tell the families of the people who died on 737 MAX disasters. "Don't worry - everything's going to be okay! The engineers learned from their mistakes - accountability works, you have nothing to be sad about!"
Tell the family of the person killed by a semi truck driver who showed up to work drunk or high: "Don't worry - the driver went to jail! Accountability prevented anything bad from happening!"
Accountability alone fails to prevent deadly mistakes millions of times a day; millions of mistakes are avoided daily through process, redundancy, independent review, and formal methods.
"Accountability prevents mistakes" is a comforting delusion. In reality, accountability is only marginally related to whether or not mistakes are made.
If a human messes up enough eventually they well get fired, fined or jailed. An AI will not.
A human also knows they might get punished if it messes up bad enough, which might cause it to think twice before doing something bad. For an AI there is a reward, but there is no risk.
So while both might lie, only the human will be worried that it will be found out. That makes a difference.
I hear you, but isn't the human in the loop precisely the one who should be putting their foot down and saying "no, the AI shouldn't be writing the tests to begin with", which would bring us full circle?
You say that like all humans are alike: that they all care about getting fired, fined, or jailed; that they're even considering punishment when they're making their decisions; that risk factors into decision making.
What you are describing is a hypothetical "rational person". In real life, even the most rational people you know do completely irrational things routinely.
The Therac-25 engineers were accountable. The 737 MAX engineers were accountable. Accountability is doing much less work in the safety story than you seem to think.
The real work is done by process, redundancy, independent review, formal methods. None of these inherently require someone to be penalized for making mistakes, and penalizing people for making mistakes is a demonstrably, empirically unreliable mechanism for preventing mistakes.
A lot of people are un-trainable. I've worked with people who got angry when the Instructor showed up and the instructor wouldn't just give them the completed exercises, while most of the other people just sat on their laptops checking emails ignore the training all together. I've worked with people that want training on how to use Claude because they can't figure out "just type what you want in the box, and it does it for you." I've worked with people who were amazed when I googled for answers and were like "I should do that to!" (but didn't)
People will demand training, ignore it, and continue to be a drain on the company. There will be those people out there that have one very very very specific skill and that's all they want to do; "I remove people from Active directory that have names that start with A,B or C" or "I run this Ansible Playbook someone else wrote, and that's my entire job."
Their tech positions were really the only common path to get continued raises and promotions starting from the bottom in GM for the past few decades. Most other positions only ever got hired from outside the company because internal hires expected raises for it, so to me this just looks like gm putting their tech department under the same self destructive hiring policies as the rest of the company.
I found that sort of thinking is no longer a part of corporate culture(in AU at least). As in, investing in your staff and planning for the future. Resources are meant to be used and abused for benefit and that's it, consequences be damned.
I haven’t heard of a single software job in the last decade that offers anything resembling “training”. You sink or swim based on your previous knowledge, what you can glean from the codebase and coworkers, and how skilled you are at self-teaching.
I don’t believe anyone trying to hire 20 bucks an hour for vibe coders is a serious employer. You’d make more money being a barista in some cities or a substitute teacher.
Best excuse they got to fire well paid and experienced proffessionals that have worked for them for decades and replace them with low paid new hires. But GM isn't in a great spot to be weathering any negative downstream effects. And after seeing how they treat employees for decades ill have zero sympathy for them when things go downhill.
Get a low paid entry level job at a company. Do a good job over many years, or decades. Raises, promotions etc brings your salary up to a decent level. The company can’t have that. They use AI to cut you and start over with another low paid noob.
Man, the only advice I can give people is do not sacrifice time with your loved ones for a company that doesn’t give a shit. Your kid is only going to graduate once. Those family vacations are priceless in the long run. Hell, I take time off to hang out with my dogs now and then. The job can wait.
That's the neat part, many don't need to worry about missing their kid's graduation because they can't afford neither kids nor the safety to have them (home ownership, etc).
I've been down quite a few rabbit holes like that which made me think that a lot of major 'issues' appear to be meticulously engineered to protect a certain set of interests at the expense of others.
It's like; "Damm, houses are expensive, I'm going to live in a caravan" then you realize you can't park it on your own land without council approval... Then you find out that council will never approve due to it "negatively impacting the charm of the area."
Then you become homeless and realize that you can't legally put your tent anywhere and all the camping sites in the wilderness which you used to go to as a child now charge you fees to stay there and have rangers patrolling constantly (paid for by your own tax money you used to pay). Also, you can't get a job without an address and it's a literal catch-22... Then if you lose hope and start doing drugs, bad actors (possibly sponsored by foreign states) put fentanyl in the drug supply to finish you off. Then the media fully covers it up by distracting people with slop.
People are dying and it is covered up in the most targeted, effective way imaginable... They are not only killed, they are blamed for what is system failure on the way out. "Should have gotten a job," or "Shouldn't have done drugs." And the people doing the most blaming and defending the system are passive-income shareholders who have a lot of time on their hands and sit at home all day and further rig the politics in their favour. It's cooked all the way down.
It's like the dystopian book "Brave New World" is looking pretty good by comparison to where we're heading. At least in BNW, the "savages" had a designated reserve they could escape to.
This will never stop being the stupid approach. Destroy morale, lose institutional knowledge, waste months or years getting new folks up to speed, all for skills that could be developed in house and with targeted hiring by a functioning organization.
Probably should just replace their software engineers too. I don't mean to throw shade, but automotive software is so bad I'd bet my life that Claude could do better.
Nevermind that the update cycle seems to be 6-10 months for changes like "You can now reset your radio presets directly from the radio settings menu", while bugs like temperature control resetting to max cool every start-up never get fixed.
I kept reading about how bad Android Auto was for years but we finally bought a more modern used car and I can't believe they would ship that experience to customers. I had a week where I just had to unpair and re-pair Everytime I got in the car.
I would love to read about why that stuff is the way it is from the engineers, hmm that might be a good spelunking. I really must be missing something that makes it harder than I think it really should be.
US car software, ya, I've never seen such trash in my life. This said I don't really have any complaints about my Hyundai software. It works, doesn't crash, and does what I want it too.
Well, if all the training data for auto software was trash anyway, then you know what they say:
Carbage in, carbage out.
> agent development, model engineering, AI-native workflows -- point directly at where large-enterprise demand is heading.
I don't understand these words. Does "AI-native workflow" mean vibe coding?
I am now seeing a lot of roles asking for "AI-enabled engineers". And I am not sure what that means either. I am sort of afraid to ask because the answer will probably confuse me even more. Maybe it's my understanding of what LLMs are and how they work that makes these words mean very little to me.
Yea it's vibe/agentic coding. That's what about half of the current jobs are right now. It's really sad. Saw a vibe coder job today, 20bucks an hour
Interesting. I saw contractor rates dropping here too. The ones I saw recently were at pre-pandemic levels so from 6 years ago. The large increase in cost of living from that time makes it even worse. The funny part is that the skills and experience that are being asked for are at a senior level. So I guess that would be senior level vibe coder at junior rates?
Management literally doesn't realize that you can't exactly vibe code a serious project? Maybe they just don't care and it's an experiment. Super low risk I guess, depending on the city they might pay the floor cleaner more than that.
FWIW, I interpreted the article as saying they're not looking for vibe coding, but AI model development per se:
"...In practical terms, GM is looking for people who know how to build with AI from the ground up — designing the systems, training the models, and engineering the pipelines — not just use AI as a productivity tool."
I can guarantee you GM is not training any of their own models.
I mean agentic coding is a thing but anyone could learn that in a day or week. So the idea of throwing away people that could be the most productive with ai of course make no sense. But it's big corporations, not everything has to make sense. It's most likely a dressed up pure cost saving framed in 2026-lingo.
I was going to say that sounds like short term gain for long term pain. But I'm guessing if there are any issues in the future, the government would just bail them out.
Yup thats exactly what they want
Cheaper younger people who dont think vibe coding is bad
Is this a good idea - probably not
It's a mistake. What you really want is senior engineers vibe coding.
It would be like hiring a junior to lead a team. They're the worst choice for that role.
Vibe coded self driving cars sounds very interesting.
[dead]
AI is a godsend for automotive coding.
It's really annoying how at some ASIL levels you need 100% code coverage of unit tests. With AI, all you have to do is to get your agent to generate the tests! Likewise with all the MISRA C requirements. Need your cyclomatic complexity to be less than 10? It's just one prompt away! Now your spaghetti code can easily satisfy the safety requirements with much less effort.
As someone who's worked on this kind of stuff at GM, I don't really get the exuberance in this particular space (not just the comment I'm responding to).
If you want 100% coverage, you just autogenerate the test cases. LLMs can't properly check MISRA requirements, so they're really just a layer on top existing automated checkers. Same for complexity metrics, it doesn't get merged if it violates that (or it's a vendor dependency you won't touch anyway).
If you care about the spirit of the rules, they're not that big a difference. If you don't care, there are already ways to do this. In either case they're an incremental change, not what I'd call a godsend.
I can't tell if this is sarcastic or not, but it seems insane to let the AI write the tests.
AI can't be held accountable, it shouldn't be writing the tests that determine whether car systems function correctly.
If you wouldn’t let AI run your nuclear power plant, you need to drink more of the AI kool-aid.
All these luddites out here. The cats out of the bag. Get with the program. Give the AI nukes already.
>AI can't be held accountable
I hear this all the time. Why does it matter? Punishing a human for making a mistake does not prevent mistakes, nor does it undo the harm of the mistake. A human saying "my bad, I messed up" and an AI saying "my bad, I messed up" are equally worthless, in a functional sense.
"Punishing a human for making a mistake does not prevent mistakes" This statement suggests you don't believe in some combination of neuroplasticity as a concept or the arrow of time.
Tell the families of the people who died on 737 MAX disasters. "Don't worry - everything's going to be okay! The engineers learned from their mistakes - accountability works, you have nothing to be sad about!"
Tell the family of the person killed by a semi truck driver who showed up to work drunk or high: "Don't worry - the driver went to jail! Accountability prevented anything bad from happening!"
Accountability alone fails to prevent deadly mistakes millions of times a day; millions of mistakes are avoided daily through process, redundancy, independent review, and formal methods.
"Accountability prevents mistakes" is a comforting delusion. In reality, accountability is only marginally related to whether or not mistakes are made.
If a human messes up enough eventually they well get fired, fined or jailed. An AI will not.
A human also knows they might get punished if it messes up bad enough, which might cause it to think twice before doing something bad. For an AI there is a reward, but there is no risk.
So while both might lie, only the human will be worried that it will be found out. That makes a difference.
There is a human in the loop that either prompted the agent or approved the code. So it doesn't matter if the AI is accountable or not.
I hear you, but isn't the human in the loop precisely the one who should be putting their foot down and saying "no, the AI shouldn't be writing the tests to begin with", which would bring us full circle?
You say that like all humans are alike: that they all care about getting fired, fined, or jailed; that they're even considering punishment when they're making their decisions; that risk factors into decision making.
What you are describing is a hypothetical "rational person". In real life, even the most rational people you know do completely irrational things routinely.
The Therac-25 engineers were accountable. The 737 MAX engineers were accountable. Accountability is doing much less work in the safety story than you seem to think.
The real work is done by process, redundancy, independent review, formal methods. None of these inherently require someone to be penalized for making mistakes, and penalizing people for making mistakes is a demonstrably, empirically unreliable mechanism for preventing mistakes.
They couldn't be bothered to train their staff?
A lot of people are un-trainable. I've worked with people who got angry when the Instructor showed up and the instructor wouldn't just give them the completed exercises, while most of the other people just sat on their laptops checking emails ignore the training all together. I've worked with people that want training on how to use Claude because they can't figure out "just type what you want in the box, and it does it for you." I've worked with people who were amazed when I googled for answers and were like "I should do that to!" (but didn't)
People will demand training, ignore it, and continue to be a drain on the company. There will be those people out there that have one very very very specific skill and that's all they want to do; "I remove people from Active directory that have names that start with A,B or C" or "I run this Ansible Playbook someone else wrote, and that's my entire job."
Their tech positions were really the only common path to get continued raises and promotions starting from the bottom in GM for the past few decades. Most other positions only ever got hired from outside the company because internal hires expected raises for it, so to me this just looks like gm putting their tech department under the same self destructive hiring policies as the rest of the company.
It really is time for them to die as a company.
I found that sort of thinking is no longer a part of corporate culture(in AU at least). As in, investing in your staff and planning for the future. Resources are meant to be used and abused for benefit and that's it, consequences be damned.
They hire anew to lower salaries, while also gaining staff with "stronger AI skills."
Then they hire people with even higher salaries than the original devs to clean up the God awful mess.
That's the neat part, they don't. Maybe they'll contract some group at some point, maybe not. Then later they'll ask the .gov for a big Ole bailout.
I haven’t heard of a single software job in the last decade that offers anything resembling “training”. You sink or swim based on your previous knowledge, what you can glean from the codebase and coworkers, and how skilled you are at self-teaching.
Laid off ai workers are cheaper. I saw a job posting for a vibe coder, 20 bucks an hour.
I don’t believe anyone trying to hire 20 bucks an hour for vibe coders is a serious employer. You’d make more money being a barista in some cities or a substitute teacher.
It's out there. Maybe it's not serious but it is kind of a weird thing to post on multiple job boards.
Sure, but they probably aren't living in those cities. In-person $20/hr in expensive city vs remote $20/hr in cheap cost-of-living area.
perhaps they tried that and the staff wasn't willing to be trained. have you ever worked at a company like that of that size?
[dead]
Best excuse they got to fire well paid and experienced proffessionals that have worked for them for decades and replace them with low paid new hires. But GM isn't in a great spot to be weathering any negative downstream effects. And after seeing how they treat employees for decades ill have zero sympathy for them when things go downhill.
You won't be asked, your tax dollars will just be used to bail them out, either directly or through tariffs. Because democracy.
You see we socialize the losses and capitalize the profits. Works great for everyone
> And after seeing how they treat employees for decades
To say nothing of their cars.
Get a low paid entry level job at a company. Do a good job over many years, or decades. Raises, promotions etc brings your salary up to a decent level. The company can’t have that. They use AI to cut you and start over with another low paid noob.
Man, the only advice I can give people is do not sacrifice time with your loved ones for a company that doesn’t give a shit. Your kid is only going to graduate once. Those family vacations are priceless in the long run. Hell, I take time off to hang out with my dogs now and then. The job can wait.
That's the neat part, many don't need to worry about missing their kid's graduation because they can't afford neither kids nor the safety to have them (home ownership, etc).
It really seems like that doesn't it?
I've been down quite a few rabbit holes like that which made me think that a lot of major 'issues' appear to be meticulously engineered to protect a certain set of interests at the expense of others.
It's like; "Damm, houses are expensive, I'm going to live in a caravan" then you realize you can't park it on your own land without council approval... Then you find out that council will never approve due to it "negatively impacting the charm of the area."
Then you become homeless and realize that you can't legally put your tent anywhere and all the camping sites in the wilderness which you used to go to as a child now charge you fees to stay there and have rangers patrolling constantly (paid for by your own tax money you used to pay). Also, you can't get a job without an address and it's a literal catch-22... Then if you lose hope and start doing drugs, bad actors (possibly sponsored by foreign states) put fentanyl in the drug supply to finish you off. Then the media fully covers it up by distracting people with slop.
People are dying and it is covered up in the most targeted, effective way imaginable... They are not only killed, they are blamed for what is system failure on the way out. "Should have gotten a job," or "Shouldn't have done drugs." And the people doing the most blaming and defending the system are passive-income shareholders who have a lot of time on their hands and sit at home all day and further rig the politics in their favour. It's cooked all the way down.
It's like the dystopian book "Brave New World" is looking pretty good by comparison to where we're heading. At least in BNW, the "savages" had a designated reserve they could escape to.
GM needs to focus on making better, safer, cheaper cars.
"Cheaper" is indeed what you get when you reduce payroll.
Shameless investor signalling.
I would like one vibecoded airbag please
Unless they plan on scooping up AI researchers, I legitimately don't understand what "stronger AI skills" is even supposed to mean.
People who can use Claude and its peers effectively. AI skills = prompting and slop troubleshooting.
No AI No Life...
Seems like there will be a small elite that wants to go up to a space station and shit on top of all the citizens below :D.
This will never stop being the stupid approach. Destroy morale, lose institutional knowledge, waste months or years getting new folks up to speed, all for skills that could be developed in house and with targeted hiring by a functioning organization.
This is only stupid if you value quality of outcome. If your sole metric is make stock price go up long enough to cash out this shit is brilliant.
Ya, we burned the planet to the ground, but for a few short moments we created a lot of shareholder value.
Well, the US auto execs visited China and realized that their days are numbered. So why not extract as much profit as possible?
Firing people with institutional knowledge? So what? It's going to improve profits short-term.
Why can’t US auto makers automate like China did?