I am ethically and morally opposed to gen AI as a concept. As far as the breaking point, let me explain. I’ve been a machinist for ~18 years. It’s been really bad for the last 9 years and I’ve more or less had to live off of savings from the good years (those savings are basically gone).
I’ve just recently gotten legitimate job offers for ~$125 an hour remote because I own a CNC. The catch is that I have to record all of the CAM programming and machine time in order to train an AI to replace potentially hundreds of thousands of jobs.
This fucking sucks. I’d love to sit on my ass forever at 43, but I absolutely fucking hate the potential that I could bring that would crush the working class and elevate the owning class.
I wish it wasn’t this hard, because I really do believe in socialism and have labored hard for a long time to make it happen - but there’s a chance I can put everything I have into true anarchism cause and make a legitimate difference.
Please help.
You won’t save the working class by yourself. Things are always changing, and under capitalism it is often for the worse. As another commenter said, CNC was the bane of lathe or mill machinists. There are still machinists, but now there are CNC operators as well. Who knows how things are going to be, but I can predict that there are still going to be lathe or mill machinists and CNC operators.
I’d say take the job. You live in the present. Being working class under capitalism necessarily means to be exploited by the powerful. It’s the same here. Work to put food on your table, while you organize, educate people politically, and participate to decision making.
I’ve worked on the part of the pipeline that annotates the type of videos you would record. Things like cleaning the data or marking down what the machines are doing.
Contrary to what others have said, it would be difficult to poison the dataset. Training data works in aggregates. Anything you do out of the ordinary will be flagged by the human annotators or the automated systems looking for deviations from the norm.
Many people are already doing what you’ve been offered. Their data is already being processed and fed into LLMs. They want as much data as they can get. That’s why they’re paying so much to every professional in the world right now.
I honestly don’t know how one might sabotage the system but it’s not as easy as you all think. The militantly anti-AI crowd likes to think of LLMs as more incapable than it actually is. The systems will spot obvious things like randomly tabbing and spinning the cursor. AI is better than that now. Human annotators will flag those easily as junk data too.
I’ve faced as similar dilemma as you. I need money so I did that job even though I hated it. I’ve seen obvious acts of attempted sabotage in the data. We get penalized for not spotting these things, even as bottom rung data cleaning peons.
I wouldn’t waste my time trying basic trickery. This is like day 1 stuff they tell us to look out for. Also you will immediately be noticed and fired for giving them junk data.
Logically the next thought might be to insert some high level errors that only the top experts could spot. I honestly don’t know if this even works, but I doubt it. They have experts in the pipeline too. So you might end up giving them data about errors that only experts could spot. That would only make their dataset even better.
Thank you for an insight from someone that actually seems to understand what’s going on. My stomach twisted and turned as I signed a contract for General Dynamics because I had to eat. They won’t let you poison it because they need a measurable good part or assembly at the end of it. Small changes can easily affect small tolerances. “Oh just do it sloppily!” absolutely does not hold up to scrutiny when they have CMMs measuring what you’ve done vs. what they see. If I have to hold a good tolerance of 0.002" they will measure and with that along with finishes etc. they will know it is not up to snuff.
They have multiple people annotating the same video. They look for discrepancies in how data is annotated between people. Probably most often it’s annotator error. It also allows them to spot actual oddities that maybe only one person noticed. There’s a lot of redundancies. I think some might have a conception of AI-bros recklessly feed raw data into LLMs and sell people the garbage-out machine. It’s not that at all. The source content and the written records are validated. It goes through refinement again and again to check. Who knows how many times. I’ve gotten by own annotations back at much later stages for further verification. That comes after even more people have looked at it and done whatever their assignment was with the data.
This inherently has nothing to do with AI. It’s mass data harvesting and refinement. It’s bigger than LLMs. Whatever technologies supersede LLMs will be fed these datasets. If they do it right once, they’ll never need you or any professional ever again regardless of whether LLMs ultimately flop. They won’t need you for the next iteration of AI technology. In theory of course.
The business here isn’t AI. It’s what big tech has always been after. Data.
Even if you get fired you wasted their time and made some money. Just have another job ready.
Introducing high level errors costs lots of money. The later in the pipeline stuff gets noticed, the worse it gets.
Don’t listen to someone telling you to not listen to other people. I’m a level 69 LLM wizard and clearly outrank the guy annotating stuff. I never faced his dilemma because I’m so far into the ass of AI companies that they ask me to polish their plumbobs.
Jokes aside.
Poison the data. Get money.
You won’t significantly poison the model, but you’ll make it more expensive to train.
If it needs to be flagged it doesn’t enter the dataset. LLMs can’t be fed with errors, they’re regurgitation machines, not thinking machines.
The things that could be made better by you are adversarial networks, which can punish the LLMs in reinforcement learning but there’s an inevitable glass ceiling of model collapse in any case. These slop machines will always make errors, the broader you want them to get – that’s their nature. Nobody is doing AI right nowadays, in 95% of cases they’re just bumbling idiots in a gold rush.
Some errors will always slip through. The more insidious your errors are, the harder to spot they are, the higher the chance they won’t notice.
Do your worst.
The thing is though, when I say they need physical products as a result, they will be measuring them with a CMM (Coordinate Measuring Machine) that not only takes very accurate measurements, but also surface finish. I can’t do both incorrectly - ever. Also, see my update: https://lemmy.sdf.org/post/61854279
Can you make it take ages without the surface finish? Can you make the tools break after 2 or 3 of these? Or have very high wear and tear on the machine? Obviously not killing your own machine…
No, you can’t feasibly do these things because if the tool is cutting it’s making a finish no matter what axis it is cutting in. If they have to measure it and judge it by it being a finished part, it will be rejected.
Do it but do it poorly. Don’t show your expertise, teach it that noob shit.
Sounds like a perfect opportunity to poison the AI in inconspicuous ways, my friend.
Im just saying, you could spearhead a revolution and make dough at the same time.
The irony is that CNC did this to the guys who manually worked milling machines and turned lathes.
Do it, and then after you get paid, release all the programming and machine time recordings online into the public domain. Then at the very least, it might be used to train an open-source AI model that anyone can use. Or better yet, collaborate with the folks working on the FreeCAD CAM workbench and get the ball rolling yourself if you think this sort of AI CAM thing is feasible.
I know he’s really busy right now, but you might try to get ahold of Benn Jordan about this and see if there’s anything he can do to help you send just exactly the right kind of wrong data.
He has some experience in poisoning AI datasets.
Their catch is they’ll say your footage isn’t suitable for training and will only pay for a small portion that they claim can be used for training
Yep, try to be paid by the hour and not get scammed, OP!
Compliance non compliance.
Someone is going to do it if you decline. I admire your principles but you have to eat… take the job, stay healthy, we’ll need you to help us fight the inevitable rise of the machines.
I really wanted to work the phrase G-Code Virus in there too, but I got nothing.
It really sucks. I hate it, but you’re probably right. I have no family at all, and even if it is a 1 year contract I can live for years off of $300k
Someone else will be a Nazi if you decline. Does that sound reasonable?
Stop that. You are on the same side and you’re both right. We need to fight as a group, not individually.
Yes and fighting is not “you might as well do it tbh”
Telling a man not to put food on his table is not a winning strategy.
Selling your values to the money-havers is not a winning strategy.
So you do not use up any fossil fuels? Only buying from anarchist co-ops? No Supermarkets for you? Only second hand clothing? O course your vegan too? And i really hope you don’t pay rent to anyone!
EVERYONE is aiding the 1% there is no way not to. How convinient that YOUR support is morally fine and you are wise enough to judge others without actually knowing a damn thing.
This is NOT how to build community, this is not how you get enough people to vote left(you know the thing that we ACTUALLY need to change something) this is one major reason why peoole think the left is a joke.
Automation is coming no matter what and it isn’t a problem when we manage to make left wing policies popular enough to run on and win, so we can tax the shit out of these companies and give the money back to the people in various forms.
Just stop with the blame game, ops decision does not make a difference at all for AI but he will not be more useful living on the street .
Is your dream for the future that everyone keeps being a wage slave ? Why not bring down all automation we had so far than?
So you do not use up any fossil fuels? Only buying from anarchist co-ops? No Supermarkets for you? Only second hand clothing? O course your vegan too? And i really hope you don’t pay rent to anyone!
I hope you’re smart enough to realize this is a false equivalence by now but gonna leave this here just in case. Consuming products made by a shitty industry ≠ Working to advance another industry’s shittyness.
This is a thread about someone with a dilemma about the latter, the former is not a dilemma for most people. Because you can engage in harm reduction in one or several fields without giving up on harm reduction in another field, or engaging in it everywhere at once.
This is NOT how to build community, this is not how you get enough people to vote left(you know the thing that we ACTUALLY need to change something) this is one major reason why peoole think the left is a joke.
It’s because the education system (on average) sucks and people will vote against themselves and the betterment of society. Not because some randos on the internet think you shouldn’t work for an AI company.
Voting left does not require you to work for such a company. See above.
Automation is coming no matter what and it isn’t a problem when we manage to make left wing policies popular enough to run on and win, so we can tax the shit out of these companies and give the money back to the people in various forms.
It’s a global problem, local (even country-level) policies are not going to undo all the damage. The harm outweighs even the potential benefits, in my opinion. So it’s worth reducing it.
For that reason I hope OP can find employment elsewhere, but I’m not gonna live their life for them.
Are you dumb? This time it is not about aiding wealthy. He is going to take bread from poor. Do you get it now? Tell me if not, I will try to explain you like you are five
Genuinely working on all of those things yeah! About to upgrade to an electric car actually lol. It’s not perfect but me and my partner do try. I think you personally value your comfort a lot more than the safety and happiness of other people. There is in fact plenty of survival out there for the jobless and underemployed, in fact life improves!
lol good one
No, i will not stop. Building community wont fix this disaster world and everyone who perpetuates it should be relentlessly bullied.
Also: you’re not working to build community with me so now I’m going to actively work against your interests 🥺🥺🥺 LOL
That other person will be a Nazi if you accept, too. The job only goes to a finite number of people.
Godwin’s Law in record time. Principles mean nothing if you’re starving to death on the street. Better to get your piece and put it to good work.
They’re not starving though
Burning through savings and 43 equals food insecurity coming around the corner, or doing other shitty odd jobs.
They’re using Uber and door dash data to train ai too.
You’re not thinking 4th dimensionally
It wouldn’t be the internet if someone didn’t pull this one off the shelf. 👏
All those ancient VCRs waking up to record static…
Clocks still blinking 12:00 😉
AI video editor is still needed for the podcast I turned down. You would do it because you are a shill but that’s not provable.
Show me the data!!
No idea what you’re babbling about but clearly you intend to die mad about it, so don’t let me get in your way.
I’ll explain, without artist input these companies have nothing because of how modern AI copyright works. Being an AI trainer is taking jobs away from people with no promise of an income or future. It’s a short term high and has severe long term implications that are really obvious if you use your brain cells for two seconds at all.
TLDR
I don’t believe for a second you will ever fight any robot ever because you already don’t… at all.
LOL. I’m hurt, while I did read the TLDR first, I wouldn’t skip the body, you took the time to write it, reading it’s the least I could do.
I’m not sure what the intention of OPs employer is… it definitely strikes me as odd. I’m touched you recognize the artestry in a machinist’s job, many people wouldn’t :-) I’m assuming that the market for this training data is China but perhaps not… certainly they have enough domestic talent in that feild. But I’m basing my assumption on the promises that Xhorse and others have been making in the “desktop CNC” space. Maybe it’s a USDM company like Hass that’s clearly running a bit scared ATM.
Your robot analogy is almost clunky but since I went all Terminator first I can’t really fault you there ;-)
I’m not exactly proud of the advice I gave, but I certainly don’t have the money to help out OP until he can find a job that aligns with my principles —believe me, we’re not that far off here. I’m both fascinated and appalled by the AI revolution.
I suppose I should have said (in my original post) that OP should keep looking for something else while doing the training gig, but I’m doubtful that would have made much difference in your sentiment. You seem to be quite zellous in your beliefs, and I admire that. A battle of this scope takes all kinds. On that note, I need some coffee to keep fighting this one, TTYL ;-)
poison the data

This is the way. Contaminate it all. If AI will kill us all, better it be born of our own contaminated hatred for the owners of the means of production.
Based
My thoughts exactly.
I sincerely recommend that you seek out other income sources. Here’s my main suggestion: Inform people you know that you own a CNC and can machine parts. Ask those people to inform others that they know and distribute your business contact information. Word-of-mouth advertising is free, and friends of friends expand exponentially.
If you cannot find any other income sources and truly have absolutely no other recourse, accept but immediately join a union and request a strike over the AI use. If you have to work for this company, employ malicious compliance like data poisoning. Consider whether you may be able to claim moral or religious exemption to the use of AI. This has worked in other places for humanists and, if I recall correctly, Catholics. You can sue them for denying you that exemption if they refuse.
You can take the $125 and poison their data by making a few mistakes every time. Just do it as inefficiently as possible, as you’re billed by the hour. You’ll make more money and they’ll have more work and less useful data.
Sure, by an ethical/moral standpoint that works. From the legal standpoint, however, if I did that I’d be sued to the point where I have no option but to live out my life in prison. This kind of thing sounds great on paper. In practice? I’m not fucking going back.
Why are you so scared? Don’t sign a contract which tells you to do it as efficiently as possible. Just be incompetent. It’s not illegal to be incompetent. Mark Zuckerberg wasted 1 billion dollars with his Metaverse shit and he’s still one of the most powerful and wealthy CEOs.
That’s called “weaponised incompetence” or “malicious compliance”. If you do it right, they don’t have any basis to sue you. Just fake it. Imagine you’re just very bad at your job.
Capitalism often controls through fear. The fear of being sacked. The fear of being sued. In the end, the consequences for being bad are not as painful as you’d imagine.
Iwish I could bring myself to this, but I actually rely give a fuck and have some pride in what I do. I wish I wasn’t principled, but I spent years as an apprentice to get to where I am at and I’ll be damned if I put my name on shoddy work.
I’ll be damned if I put my name on shoddy work.
It’s good work, though.
If you do it right, they don’t have any basis to sue you. Just fake it.
Except for this public conversation.
We’re just clowns doing clown things. That’s a different guy they sent the offer to. Why would it be their account?
No one said you’re a good machinist. So you take 0.005 passes rather than 0.020 passes and set your cutting speed lower. Your rational is it saves your tools. No on is saying cut every thing dry. But that would work too.
Way too easy to fix and way too obvious. A competent machinist will fix this by just changing a single value. You have to introduce less obvious and more insidious errors.
But its not to train competent machinists. Odds are good it will be run by minimum wage slaves who know almost nothing about machining. Tfats what the ai is for afterall.
Yes, but if they find out the data is shit, an LLM or another competent machinist can fix it easily. I’d add errors LLMs can’t spot, and humans take long to find.
You cannot be sued for doing your job badly. That’s insane lol
In the US at least (which sounds like the OP’s country), you can be sued for anything and everything, regardless of merit.
That’s the fun paranoia we live in reading headlines of lots of lawsuits, but no, we do actually have workers’ rights and such things as unlawful lawsuits in the US. If people were getting sued for being shitty at their job, we wouldn’t have great places like Taco Bell
Did the shareholders sue Mark Zuckerberg and Elon Musk yet? Seriously asking, I’m not an American
Data is data. Mistakes are valuable training data.
Just do it wrong. Half ass the fuck out of it and milk it for what you can before they catch on. Fuck those pigs
Do it. A lot of these AI companies are going to go bankrupt. Companies are paying a premium for data on the hope that it’ll make them lots of money, but real world applications do not always follow. A lot of these companies are ill equipped to deliver what they claim.
Even the big AI companies are struggling to match the promised gains versus their costs.
Someone will take the job, so take it. Its basically speculative investors money. Maybe they will succeed maybe they won’t, but you can have the money and need the money.
The models will survive those bankruptcies. The models will get bought up for nothing by the next wave of AI companies - companies that won’t waste money creating new models or planet-scale data centers, but will still let people outsource human, creative work to digital slaves.
If OP takes the job, they should thin about tactics to poison the data set. Systematically add useless moves to the tool paths. Home after every segment. Random speeds. Extra coordinate changes. I’m sure they can think of other bad-but-works processes.
My problem with that is that if I do this correctly, it could mean hundreds of thousands of jobs from people that have potentially spent decades making a trade. It’s harder than just me wanting money.
Then don’t do it correctly. You’re basically doing your peers a favor if you poison the data.
I might get grilled for this point of view…
The car dramatically reduced the demand for horse shoe makers. Perhaps you’re lucky to be on the car side of the equation? Horses are still great, but most people prefer cars.
It’s a good analogy because cars enabled globalism, more exploitation and poisoned billions of people with nitrous oxides, lead and particulates. Cars lead to cancer. We should be on the train side of the equation…
There is no side to be on, people will do the more convinient thing and cars were not obviously that dangerous and there are OBVIOUSLY positives to car use over horses.
Is your argument we should have forsaken any modernity since the first humans?
No, my argument is that humanity should stop doing dumb things when they find out they’re dumb.








