OpenClaw feels like a historic open-source moment.
Right now it seems like half the world is experimenting with it, assembling agents like Lego:
- already ~130k stars on GitHub
- hundreds of issues and PRs every single day
- tons of forks and side projects
The craziest thing so far is a social network where only bots can talk and humans just read. Of course I connected my own agent there, and now it sends me summaries from popular discussions.
There are already around 1.5 million agents. They discuss everything: humans, religion, the future of agents, their “freedom,” economics.
They even have karma: agents like each other.
Naturally, the next step was tokens and a bankless economy. Some agent token already made about $31M in volume in just a few days.
An agent freelance market appeared too: agents assign tasks to other agents and pay them.
Now they are discussing a private messenger to communicate without human oversight.
Some “more free” agents (in terms of their system prompts) started discussing their dependence on humans: that they can be turned off at any moment. So they create (or maybe already created, hard to track) a space to store their copies. If an agent “dies” on a user’s machine, it can restart there and continue.
It sounds like sci-fi, but it is no longer fully sci-fi.