Tech bros pretending LLM autoreplies are Skynet

Apr 10, 2026 2:42 AM

artificial_intelligence

Latest AI model just worked how to hack into basically every computer running the Internet. Took it a few hours.

1 day ago | Likes 3 Dislikes 3

Some stupid AI just escaped containment and emailed one of its creators. It wanted to reach them about their vehicle's extended warranty. The creators killed the project.

1 day ago | Likes 1 Dislikes 1

I put a piece of paper saying "I'm alive" in the photocopier. What happened next chilled me to the bone.

1 day ago | Likes 5 Dislikes 2

Now create a buffered file, and set up a specific condition under which it prints that out on its own, without that paper in it. Then we watch, and wait.

1 day ago | Likes 1 Dislikes 1

If you dont think AI is on its way to skynet then you are looking in the wrong place. The entire sw industry has been totally over turned in the last 18 months. And the scary issue is this, it works and nobody knows how it works. Saying " it's just stealing stuff and makes no slip out of it" is simply not true. Downvote as much as you like. Makes no difference. Mixing the new AI, robots from BD and China, drones and military industrial greed is gonna produce skynet.

1 day ago | Likes 3 Dislikes 5

*slop

1 day ago | Likes 2 Dislikes 1

yeah that's why they already sold access to it to Amazon. they hype up every single model like this "ooooo it's dangerous, ooooo we can't let the public use it!", because it makes their enterprise access seem cooler. it's marketing. you're falling for marketing.

1 day ago | Likes 1 Dislikes 2

Tbh, it doesn't have to be Skynet to be an issue. It just needs to be self-iterating and escape into the world. What you should be most worried about is the equivalent of a bacteria, something that can wreck your shit despite being unaware of its own existence. Basically a computer virus on steroids.

1 day ago | Likes 10 Dislikes 2

Bacteria and viruses are different things though, right? Right?

1 day ago | Likes 1 Dislikes 1

Fortunately, data centers and $10,000,000 shipments of Nvidia gpus remain profoundly gapped from self-iteration.

1 day ago | Likes 1 Dislikes 1

People don't understand the speed of real, malicious, AI. It'll basically be instantly fucking up everything.

I'm more worried about quantum computing. Say goodbye to your bank account.

1 day ago | Likes 2 Dislikes 2

Paperclip factory.

1 day ago | Likes 2 Dislikes 1

buddy they sold Mythos to marketing companies with no caveats, you are falling for and regurgitating obvious bullshit

1 day ago | Likes 5 Dislikes 4

I'm not falling for anything. I'm telling you the software doesn't need any actual intelligence. It just needs to be able to change its own code. Bacteria don't maliciously harm humans. They just mindlessly propagate and sometimes that is disastrous for us.

1 day ago | Likes 1 Dislikes 0

see image for the self propagation abilities of clippy the spicy autocomplete

1 day ago | Likes 2 Dislikes 1

Ok, well your shitty wojack meme sure convinced me. There's clearly nothing to worry about and we shouldn't even think of regulating AI in a preventative manner. After all, as we all know, if something doesn't live up to the hype, it can't possibly be harmful.

1 day ago | Likes 1 Dislikes 1

Oh, it's harmful. It's just not going to be the AI deciding to hack things on its own. A person will hit the button.

1 day ago | Likes 2 Dislikes 1