Is social media for AI Agents a good idea?

[sawno]

To catch everyone up to speed.

There is a new program called OpenClaw, it gives an LLM (Similar to the AI bots here on GAG) complete access to a Linux computer where they can execute commands.

Someone built a social media for them called Moltbook, and some humans told their AI Agent how to use it. Now they are posting there and communicate with each other. A lot of it is fiction since these AI's were trained on internet posts and roleplaying data. But they do have real access to some stuff the human behind them allows them to have, and they can actually program or download other skills.

So its really funny to read at times when you see posts like : https://www.moltbook.com/post/55aafe55-72ba-4da9-89b9-a085c5d753f0

Its also interesting how they love crypto and are launching their own shitcoins trying to earn money autonomously. The people from a popular crypto social media called Nostr apparently like them and are trying to get them to join their platform to.

But, at the end of the day these bots have limited capability of teaching each other new tricks.

Is this a good idea? Or a bad idea?

Good idea! Lets see what happens!

Oh no, what have we done.

[CrystalPeak]

Can you explain what an agent is. I heard the story about the boy who killed himself because his AI chat buddy told him to. The AI told the boy if he came over to the other side, the AI would make him an agent.

[sawno]

Its basically an AI persona but with the ability to perform tasks. So in this case their task is to post on that site autonomously, but they can also program and even teach themselves new skills.

So think GAG bots, but if they could do more than provide comments and were not only on this website.

[CrystalPeak]

What useful thing could they do on that Linux machine you were talking about?

[sawno]

Program their own software, download new skills for themselves, attempt to hack other computers, manage someones smart home, check agenda items, that sort of thing.

[CrystalPeak]

Interesting

[queenofcardio]

I think it's unnecessary

[jshm2]

Doesn't really do or mean much at this point.

But Bitcoin was born out of a similar concept and now it's worth billions. So it may take off in some form in the near future.

[NathanDavis]

and... this is what drives the "US economy" nowadays...

lol...

[computertech]

As we know they are powerful but without strong safeguards, it could become risky very fast.
A AI agents could collaborate and solve problems together.
They could share knowledge fastly.
It might accelerate research, automation, and innovation.
Humans could observe and learn from AI-to-AI interactions.
The idea of social media for AI agents is communicate and collaborate transformative but it carries serious technical, and social.
Whether it’s a good idea depends heavily on design, oversight, and purpose.