this post was submitted on 12 Nov 2026
276 points (97.9% liked)
Fuck AI
8779 readers
1102 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
Rules:
- Relevant posts only.
- No bigotry.
- No misinformation.
- No NSFW/NSFL content without being labeled NSFW or being in an appropriately-labeled spoiler tag.
- This isn't a place to talk about the cons and (supposed) pros of AI for different use cases. This is /c/fuck_ai .
- Especially no AI advocacy.
- No AI-generated content. If you want to post examples of AI failing hillariously, there are other communities for that like !ArtificialIgnorance@lemmy.ca .
- Venting is part of the purpose of this community, but don't aim frustration or accusations at other users.
- These rules are a living document that may change in the future.
You can help the mod team by making your reports detailed.
Effective reports:
- "The third image is a well-known anti-semitic trope"
- "Claiming that 'agentic AI' can outperform humans at all jobs is misinformation"
Less effective reports:
- "Breaks rules"
- "Misinformation"
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Surely there are already generic CNC emulators? I had an internship 15 years ago for a CNC company that had an emulator for their tools. I also know it was leaked on the pirate bay, which everyone thought was hilarious. Pirating the emulator is worthless without the very expensive CNC machine.
I really doubt there is anything an AI can learn from your videos, and I suspect the AI company has no idea what they are doing. Take the job, get the money (in cash), and wait for them to crater. Dont be surprised if the job dries up long before the term though.
I ran a hard mill at one point and occasionally a cnc lathe. That was 20 years ago and there was a ton of automation. In fact, it was mostly automation. I'd indicate the die or punch into the machine, load the igis file, and then go sweep the floor for the next 8 hours. There were a few things that automation couldnt handle (it would leave sharp edges, for example), but I think most of that was the fault of the dude that built the igis file in CAD. So Id be curious as to what AI would do that it wasnt already doing way back then.
Anyhow, didn't love that job and left quickly.
When I worked my CNC job there was lots of talk of light out manufacturing. Factories with busy machines, and no humans, so no need for lighting for the useless squishy humans. Its always been something that engineering has been working towards, it was just a lot more gradual until recently.
There are generic CNC emulators. They've been around for a long time (like 15 years, as you said). The difference is that that they will do the grunt work, but a human needs to make the tool paths actually viable. The difference is that with actual professional tools like MasterCAM, it has been cutting out that intern for just as long but a cost of thousands of dollars per seat. What this has the potential to do, if I do it correctly, is even push out those incredibly expensive MasterCAM seats. I could cost hundreds of thousands of jobs if I do this correctly.
The emulator in had could do tool path correction, and tell you when you would crash. And they were working on lights out automation when I left, albeit classical programming, not AI. Was easier for us given we only supported a very limited number of machine models though.
I suspect, and you'll know better, but these guys dont know shit about machining, which is why they are coming to you. I suspect that they have no idea how large the problem space actually is, and they are just going to try to brute force an AI solution.
And I dont think that is viable for machining.
Whether you do this correctly or not won't change that they dont know what they are doing. My guess is you can give them exactly what they ask for, and after a month or 2 it won't work, and theyll ask you to redo it all slightly differently, and youll go through that cycle until they run out of money.
You may as well take their money. And be smart about it. You need a top end dSLR camera to record accurate footage, get them to send you one. You need a top end PC and a high res and high framerate monitor to process the footage.
Do you think you could do it in a way such that you introduce suboptimal patterns into the training data, to make the model less viable and less reliable? Because poisoning training data is something that a lot of people are starting to do in a lot of domains.
I don't think so, because they want a viable physical product which can be measured accurately.
Make it twice, record the bad one, ship the good one.
.. Until ArtCraft clones it for free?