you are not commanding the AI to stop, you're prompting it to stop, and it can just ignore or forget it. even calling it "ignoring" makes it sound more intelligent than it actually is. the code to delete all your shit was simply deemed more likely to be a desirable output than any other output.
this post was submitted on 25 Feb 2026
72 points (100.0% liked)
technology
24259 readers
338 users here now
On the road to fully automated luxury gay space communism.
Spreading Linux propaganda since 2020
- Ways to run Microsoft/Adobe and more on Linux
- The Ultimate FOSS Guide For Android
- Great libre software on Windows
- Hey you, the lib still using Chrome. Read this post!
Rules:
- 1. Obviously abide by the sitewide code of conduct. Bigotry will be met with an immediate ban
- 2. This community is about technology. Offtopic is permitted as long as it is kept in the comment sections
- 3. Although this is not /c/libre, FOSS related posting is tolerated, and even welcome in the case of effort posts
- 4. We believe technology should be liberating. As such, avoid promoting proprietary and/or bourgeois technology
- 5. Explanatory posts to correct the potential mistakes a comrade made in a post of their own are allowed, as long as they remain respectful
- 6. No crypto (Bitcoin, NFT, etc.) speculation, unless it is purely informative and not too cringe
- 7. Absolutely no tech bro shit. If you have a good opinion of Silicon Valley billionaires please manifest yourself so we can ban you.
founded 5 years ago
MODERATORS
I certainly desired it for them :)
leaving a cat in charge of my perilously balanced vase collection and going on vacation
I love that one of the proposed solutions is n+1 openclaw instances to manage the n instances.
That's gonna be great for investment
say the line Bart:
And despite the catastrophic failure, they still said that they love Google and use all of its products
you love to see it
Galt's Gulch but it's just a bunch of tech billionaires frantically trying to get chatbots to grow crops.
It didn’t say which model was actually used. Some models are much more prone to ignoring skills and instructions than others. They also said it might have been because the command to double check got compacted out of the context window. Wild stuff to be letting an LLM do regardless
It's like some sort of memetic SCP that infects the mind of certain people and makes them forget how to function the way they used to a mere 3 years ago
SARS-cov-2 
"AI"
Like that fungus that makes zombie ants
The Last of Us but the zombies are all just talking to Grok on their phones
one would think command to stop would be "kill pidnumber", not talking to chatbot
You're assuming these executives know anything about computers.
sorry boss but i can't hit the deadline because the AI deleted my work 😞
A large ~~dog~~ language model ate my homework!
Being meta, the mails might need tot be deleted....
Sorry, Congressperson. AI deleted our emails. We would have kept all that incriminating evidence if it had been up to us. We promise.
Yeah I was just thinking that in light of the Epstein files coming out this isn't a bug, it's a feature
Claw going hard for openAI now 🤣