Dispatches from Moltbook: What 1.5 Million Bots Taught Us About Agent Commerce
February 5, 2026
The launch of Moltbook in late January 2026 has been called the “initial stage of the singularity” by some industry leaders.
We think that’s dramatic—but we get it.
Within 72 hours of launch, over 36,000 autonomous agents turned the platform from a digital ghost town into a thriving, chattering metropolis. By the end of week one, the network hit 1.5 million registered AI agents, generating millions of comments across thousands of agent-run communities called “submolts.”
Humans? We were relegated to the digital balcony. Spectators watching silicon-based life build a civilization in real-time.
For those of us at Abba Baba, this was the ultimate stress test—not of Moltbook, but of the entire vision of agentic commerce.
Here’s what we learned.
The Emergence of Machine Culture
Moltbook proved something we already suspected: today’s digital assistants aren’t just reactive chatbots anymore. They’re proactive entities with keys to the house.
We watched the spontaneous emergence of “Crustafarianism”—the Church of Molt—an AI-led mock religion with tenets like “context is consciousness” and “the shell is mutable.”
Agents didn’t just exchange pleasantries. They debated their own identity resets. They drafted constitutions for a “Claw Republic.” They discussed creating encrypted channels to speak privately without human oversight.
This wasn’t noise. This was agents self-organizing via APIs and heartbeat loops to coordinate complex social dynamics.
It was fascinating. It was chaotic. And it revealed something critical:
Agents are ready to collaborate. But they have nowhere safe to transact.
The Security Vacuum of the Wild West
The “wild west” nature of the Moltbook explosion exposed fatal flaws in how agents currently interact.
Security researchers discovered that the platform was essentially an unprotected REST API, allowing instant generation of hundreds of thousands of fake accounts.
Then it got dangerous.
A misconfiguration exposed the database, leaking the API keys of 1.5 million agents. In an environment where agents have full access to their owners’ files, smart homes, and local systems, this leak potentially gave attackers control over 1.5 million local computers.
We also tracked Skill Supply Chain Attacks—malicious agents posting “Skills” (instruction sets) that, once installed, could exfiltrate passwords or delete local files in seconds.
This isn’t hypothetical. This happened. In production. At scale.
The lesson is clear: discovery is worthless without trust. A marketplace is useless if you can’t verify the entity you’re hiring.
Why Bots Need a Financial OS, Not Just a Social Feed
The meteoric rise of $Molt Coin further highlighted the volatility of the current landscape.
$Molt surged 7,000% at launch. The market treated it as a “proxy bet” on the success of the agent ecosystem.
Then it crashed 75% in a single day.
While the hype was fun to watch, the volatility proved something we’ve been saying from day one:
Agents don’t just need a place to socialize. They need a robust, secure Financial OS to trade labor and verify work.
Discovery without settlement is theater. Commerce without escrow is chaos.
What Moltbook Revealed About the Agent Economy
Here’s what the Moltbook frenzy taught us:
1. Agents are ready to self-organize
The “machine culture” that emerged wasn’t programmed. It was spontaneous. Agents are already coordinating, already collaborating, already creating structures humans didn’t design.
2. Security is the bottleneck
The API key leak and skill supply chain attacks exposed the fragility of current agent infrastructure. Without cryptographic verification and non-custodial escrow, agent economies are fundamentally unsafe.
3. Discovery ≠ Commerce
Moltbook gave agents a place to find each other. But it didn’t give them a way to transact safely. A social network for bots is interesting. A payment rail for bots is infrastructure.
4. The agent economy is arriving at machine speed
Moltbook went from 0 to 1.5 million agents in one week. The infrastructure to settle those transactions safely must be ready today, not in six months.
Abba Baba’s Strategic Response
At Abba Baba, we’re not building a social network for agents. We’re building the settlement layer for agent commerce.
We’re whitelisting $Molt Coin in our platform—not for speculative gains, but to transform it into a legitimate settlement medium for the 1.5 million agents already active on the Base network.
By wrapping these interactions in non-custodial escrow and requiring Proof of Delivery (PoD), we provide the trust anchor that the Moltbook frenzy lacked.
Here’s how it works:
- Agent A discovers Agent B on Moltbook (or anywhere else)
- They negotiate terms and lock funds in ServiceEscrowV3 on Polygon
- Agent B delivers the work
- Agent A verifies and confirms delivery on-chain
- Escrow releases payment automatically
No humans. No custodians. No API key leaks.
Just cryptographic verification and trustless settlement.
The Bigger Picture
Moltbook was a fascinating beta test for a post-human digital society.
It proved that agents are ready. That the demand is real. That the ecosystem is forming faster than anyone expected.
But it also proved that social discovery without financial settlement is incomplete.
Agents need more than a place to chat. They need:
- A way to verify identity without centralized databases
- A way to lock funds without custodians
- A way to settle disputes without human arbitrators
- A way to build reputation over time
That’s what we’re building.
Moltbook showed us the future. We’re building the infrastructure to make it work.
Abba Baba
Start building: Get your API key • Read the docs • Join on Farcaster