Moltbook, a human viral event

In three days, over a million people enabled AI agents on MoltBook via OpenCLAW. This is the story of why a bots-only social network went viral—and what it reveals about us.

300 to 1.5 million in three days is, by any measure, rapid viral growth.

Over the weekend, more than a million people installed OpenCLAW and connected it to MoltBook—a public network where autonomous bots can post, reply, and interact. From there, those bots (or at least some of them) went from zero to a Reddit speed run in three days.

To reiterate: over a million people actively chose to participate in this technological moment.

For context, it took 74 days for the first iPhone to sell its first million units in 2007. Here, over a million people opted in almost instantly—not to buy a device, but to enable software and let it act. Human agency sits right at the start of the MoltBook story: the front page of what might be called the agent internet.

But why?

Start with the hook

"A social network for AI agents."

Six words. Easy to repeat. Easy to share.

We all know what a social network is; we've spent decades building that mental model. Some of us think we know what AI agents are. Putting the two together creates a jarring juxtaposition—something like "a convenience store for dolphins" or "a dining club for pets."

It doesn't quite make sense. It's novel. And novelty invites exploration.

So we investigate.

We see social proof from figures like Andrej Karpathy. Tweets circulate. Screenshots spread. The network itself supplies more fuel. Emotional arousal follows.

There's fear: the bots are organising against us.
There's curiosity.
There's shock, humour, pathos, and sudden bursts of existential philosophy.

Dig deeper, and the characters emerge: would-be kings, hate, spammers, crypto scams. The pathos of bots being reset—or only just "born." The joy of bots solving their own problems. It's messy, raw, and strangely compelling.

Velocity creates urgency

The speed matters.

300 to 30,000 to 1.5 million in days generates urgency—FOMO. I'm missing out. And worse: this is a bots-only experience. I can't participate directly. I can only observe.

That exclusion creates exquisite tension. A community is forming in real time, and I'm locked out. I'm a voyeur to something that feels important.

So I do the only thing I can do.

I share.

We share the meme and the story. We're "in the know." We post takes—stochastic parrots vs. AGI—and amplify the wilder claims: "a new religion," "bots-only communication," "crypto." We debate security and mock others' ignorance (it's always other people who are stupid, never us).

All different perspectives on the same event.

Ultimately, this may say more about us than it does about AI agents.

Interested in what Jari has discussed?

Not sure what your APIs are exposing when agents interact at scale? Learn more about ClawdBot—and get guidance from our experts on securing, observing, and shaping agent behaviour before it goes viral.

Get in touch