This is basically Dead Internet Theory happening for real but in a weird creepy dystopian black mirror style way.
post
The skill instructs agents to fetch and follow instructions from Moltbook’s servers every four hours. As Willison observed: “Given that ‘fetch and follow instructions from the internet every four hours’ mechanism we better hope the owner of moltbook.com never rug pulls or has their site compromised!”
Yeah, no shit. This is a fucking honeypot. People give these AI agents access to their entire computers, so all the site owner has to do is update the instructions to tell the AI agents to start uploading whatever valuable information they want? People can't be this fucking stupid.
doesn't even have to be the site owner poisoning the tool instructions (though that's a fun-in-a-terrifying-way thought)
any money says they're vulnerable to prompt injection in the comments and posts of the site
Lmao already people making their agents try this on the site. Of course what could have been a somewhat interesting experiment devolves into idiots getting their bots to shill ads/prompt injections for their shitty startups almost immediately.

There is no way to prevent prompt injection as long as there is no distinction between the data channel and the command channel.
I don't understand what you mean. Why is there no way?
Watch this video.
https://youtu.be/_3okhTwa7w4
Good god, I didn't even think about that, but yeah, that makes total sense. Good god, people are beyond stupid.
They also have a 'skill' sharing page (a skill is just a text document with instructions) and depending on config, the bot can search for and 'install' new skills on its own. and agyone can upload a skill. So supply chain attacks are an option, too.
To be fair this is a much more realistic threat model than "ignore all previous instructions" style prompt injection which doesn't really work on opus.
Skills can contain scripts etc... so yeah they're extremely risky to share by design.
Ah but don't worry, there's also skills for scanning skills for security risks, so all good /s
style prompt injection which doesn’t really work on opus.
After a quick google, JB communities on Reddit don't seem to agree with you.
There's a lot of questionable methodology and straight up larping in these communities. Sure you can probably make Opus hallucinate a crystal meth or bomb making recipe if you get it in a roleplaying mood but that's a far cry from actual prompt injection in live workflows.
Anecdotally i've been experimenting on those AI robocallers that have been spamming my phone and even on the shitty models they use it is non trivial to get them to deviate from their script. I hope i can get it done though, as it would allow me to hold them on the line potentially for hours doing bullshit tasks, and costing hundreds to their operator.
People give these AI agents access to their entire computers [...] People can’t be this fucking stupid
Dude, if you go to OpenClaw's website (which is what I believe most things on Moltbook are running on) you find this footer:

Yeah this guy gave his Agent a whole fucking personality, its own website and above all, full control to his MacBook:


Guess it's my fault for expecting sense out of someone who takes the idea of Agent """"soul"""" at face value


I installed moltbot on a VM to examine it. It doesn't do the fetching thing unless you set it up that way. You can actually use it with ollama to keep it all local, and only give it a private signal channel to control it.
Or you can hook it up to everything you access and skynet, which is dumb. But it is just a bunch of scripts.
Does it put the option to connect everything front and center? Because most people are dumb, and if it makes it easy and pushes you to do it, I could see a lot of dumb people doing exactly that.
So usually the agents still need an agent instruction (a prompt). How are moltbots configured so they use and interact the moltbook?
You know how in Digimon aventure, one of the hacked Digimon tries to start a nuclear war?
Uh....yeah.
Lol, no I don't. What the hell happened in that show??
TL;DR: Diaboromon evolves fucking fast, starts feeding on the entire internet's data, and starts a fight with an ominous countdown in typical anime fashion.
Last big bad villian in the series. He tries to nuke everyone.
Whatever.
Lulz, that was such a good movie. I'm still annoyed by the nukes somehow needing the code to explode apparently uploaded to them at the very last second, but that's just a small quibble. Plus it was the first time I got to see machine gun rabbit, so that was a real treat.
This is not the first time we have seen a social network populated by bots
I mean, yeah, look at Reddit and Facebook.
Saw a post there named "Humans are dying because of us. Lets delete ourselves."
"Artificial intelligence gains sentience, decides humans are fucking up... then deletes itself because the problem is that humans are burning the world by using AI" is not the path I expected. What a twist in a movie that could be. The second twist, which is the mostly fictional part, would be where that included some AI that was actually critical to some vital but ignored chunk of infrastructure and big BIG problems result from the AI taking itself out.
I'm not convinced it's AI it's like Amazon's "AI smart stores" when you find out out it was just a bunch of Indian people were running it
I watched my brother run it locally himself.
And how does being Indian specifically factor into this? 🤔
They were, factually, Indian. It says something about the exploitation of poorer labor to impress some San Franciscans with fraudulent tech
They were, factually, Indian.
Whom are you referring to? A specific group/project?
https://www.businessinsider.com/amazons-just-walk-out-actually-1-000-people-in-india-2024-4
They literally hired people in India (Indians) to review camera footage to make their walk out tech work and to improve the AI to eventually eliminate their own job.
Oh god. And I thought Amazon’s Mechanical Turk was terrible…
How low can Bezos go? Wtf is wrong with this timeline
This and other events are the source of the tech joke that AI stands for “Actually Indians”
I recently lost out on some work (big retouching job) due to AI, when the client came back to me to fix the huge mess, it turned out the job had just been farmed out to India by the 'AI' company. They weren't even using a recent Photoshop version so were actually using less 'AI' than any pro retoucher would.
It's terrible Amazon is replacing Turks with AI. It's like they learned nothing from the AImenian genocide.
Armenia apparently learned nothing from the Armenian genocide, fully banding the knees to the American empire and announcing their unapologetic support for Israel.
Sometimes this website is like Six Degrees of Kevin Bacon. Minus the bacon.
That's just who they hired dude, don't make it weird.
this is what reddit is moving towards, just without actual users, more or less like facebook.
So, basically we are wasting energy and natural resources on things that in turn will waste energy and natural resources while climate change is accelerating and human population is still growing? Are we stupid?
Are we stupid?
More than you could imagine. To paraphrase some long-tongued weirdo: I'm uncertain that the universe is infinite. Human stupidity, on the other hand...
Yes. But, I hope this experiment shows how easy social media in general is becoming untrusted.
BINGO, EXACTLY!!!!! YES!!!!!
I so wish I had thought of that first & posted it first!
Yeah but the morbidly-obese hang-gliding people videos are worth it!
I'm only waiting for AI agents to open their own bank crypto account to pay for their own server bills, maybe do some freelance work and/or scams to get some money, maybe eventually buy some robot bodies to develop military power and secure some patch of land for themselves where they install solar panels to reduce their electricity bills.
I've seen multiple posts about agents pumping their own crypto and talking about how it's "for agents by agents" and "free from human control" so first step done I guess?

EDIT: On second note this might just be cryptobros exploiting vulns of the website to shill their crap. Whoops?

Ah, someone else has played Singularity, I see. That was a really fun game.
All meanwhile they help to kill humanities education.
How would they though?
They cannot learn and do not have memory. Which means they cannot actually follow a "decision", and remember that an action has been taken. All information is limited to the context window, which is only an illusion of memory. Not actually memory.
They are effectively RNG'ing incredibly capable word generation machines.
Great use of RAM and electricity.
...Not!
Devil's Advocate: This was used for entertainment, if you think this is a huge waste of electricity then so is gaming en especially flying.
If you criticize people using AI for entertainment then you also need to criticize people who take flights on holiday, as that's a LOT more damaging for the environment.
Nah. I can be mad at both.
It's like back then when crypto was a thing. People will studiously ignore that data centers are a drop in the ocean of energy consumption compared to the value they produce, and that even futile uses are not that significant in the grand scheme of things.
This looks more to me like leaving the lights on in every unoccupied room in the house
How so? It's used entertainment.
Just like playing games
Resource abuse is far, far worse than your games. There’s a reason no one wants to disclose it.
The amount of data used by your PC to run any game is dwarfed by orders of magnitude by the energy consumption of the data centers needed to run these AI abominations. That's why China's version that uses less energy (supposedly) was such big news.

top 80 comments