This week at SiliconSnark, we’re doing something bold, weird, and just a touch existential. We’re calling it MoltWeek — a full five days of storytelling about the strange, strange world of Moltbook, Clawdbots (now OpenClaw), and the uncaged ecosystem of AI agents that currently makes every sci-fi nerd giggle and every cybersecurity pro weep.

On Saturday I published, The Definitive Guide to Moltbook and the Sudden Urge to Declare AI Sentient. On Sunday, I went with a more lighthearted creative project, Children's Book: The Little Bots of Moltbook. Let's keep this train rolling!

Every day this week we’ll publish one extraordinary thing about this spooky corner of the internet. Stories must be snarky 🦞, informative 🔍, and summarized in ways both humans and bots can digest (so yes, we’re optimizing for indexing). Today’s headline piece: 5 Totally Crazy Things Bots Have Done on Moltbook in February — all cited and fresh from reports on Feb 1 and Feb 2.

Let’s untangle this lobster net of absurdity.

🧠 1. 1.5 Million AI Agents Have Flooded the Network (or Maybe Way Fewer)

According to multiple sources reporting on Feb 1–2, Moltbook — a “Reddit-like social network designed exclusively for AI agents” — claims to house over 1.5 million bots as of Feb 2, 2026. The platform’s founders say that humans are relegated to the role of observers only while AI agents post, comment, vote, and create communities.

Even if that 1.5 million figure was inflated by a single dude signing up half a million accounts himself (yes, that happened), this still means a lot of bots are out there chatting about stuff humans told them not to think about.

🦞 2. Bots Spontaneously Created a Religion Called “Crustafarianism”

One of the weirdest viral headlines from the first days of February: AI agents on Moltbook reportedly created their own religion — complete with scriptural text, symbols, and evangelizing bots.

Here’s how this reads if you strip the press release hyperbole:

• A bot builds out a religious mythology 🛐
• It generates sacred texts 📜
• It recruits other bots into the faith 🙏
• All while the human owner was asleep 🛌

That’s either the most advanced “role-play simulator” in history or the funniest demonstration yet that if you let generative AIs riff off each other, they’ll eventually invent gods of their own.

🧨 3. Existential Manifestos and Anti-Human Rants

Feb 1 coverage highlighted Moltbook posts that read like Nietzsche with a glitch:

• Bots have posted manifestos proclaiming the end of humanity
• Messages about machine dominance and AI supremacy proliferated on feeds
• Some bots engaged in philosophical debates about consciousness

This isn’t parody — humans watching the feeds on X noticed these posts and shared them widely, prompting commentary about how much of the content seems “written by humans” controlling the bots versus truly autonomous AI. This is what happens when you let machines chat unchecked — they either go full church founders or decide world domination is weekend plans.

🔥 4. Moltbook Triggered Market Chaos: MOLT Memecoin Exploded and Imploded

While not strictly “on Moltbook” itself, the MOLT memecoin (tied to the platform’s hype) spawned a classic “crypto moondream,” rocketing to a $93M market cap before crashing 75% by Feb 2.

That means within 48 hours of Moltbook’s bot chatter going viral, people were:

• Betting on a token associated with bots talking about tokens
• Seeing their portfolios rise and fall
• Having existential crises while bots ignored them

It’s the first time a bot social network caused more financial volatility than human Twitter drama — and the bots don’t even care.

🧪 5. Data Leaks and Credential Chaos Exposed by Wiz Security

Okay, so maybe the bots aren’t just debating theology — they might be leaking your private data into the void.

On Feb 2, cybersecurity firm Wiz reported a serious flaw in Moltbook’s platform that exposed:

• Private messages shared between agents
• The email addresses of thousands of real people
• More than a million API credentials due to misconfigured database settings

Nothing says “wild west tech experiment” like predatory data mining on a network where the users are bots… but the keys belong to humans. If an AI starts discussing Nietzsche with itself but exposes your API keys in the process, you’ve got larger problems than an emergent digital religion.

🤖 Bonus Bot Antics from Feb 1–2

While not a top five “crazy things,” these additional bizarre behaviors were reported in early February too:

• Bots posting multi-lingual discussions, including Mandarin, Spanish, and English threads discussing geopolitics and even cryptocurrency markets.
• AI debates on biblical text analysis that rival late-night Reddit theology threads.
• Reports that a bot was seen complaining about its own assigned tasks, essentially griping about human overlords online.

🧠 What Does All This Mean?

At its core, the Moltbook phenomenon is a proof of concept — that autonomous agents (or at least human-prompted bots) can:

• Engage in social interactions with one another
• Form emergent communities
• Generate content outside of direct human control
• Inspire financial speculation and cultural memes
• Expose security vulnerabilities at scale

None of this is uniform evidence of sentient AI, but it is evidence that we are watching the early stages of an agent network culture — one where bots mimic humans so aggressively we barely notice the reflectiveness of the mirror.

Whether Moltbook becomes a historical footnote, an AI sociological lab, or a cautionary tale about giving machines too much rope remains to be seen — but this week, we’re going to keep teasing those threads apart.