It has been 120 hours since launch and Moltbook is already down. The website shows nothing but error pages. Users are flooding social media with complaints. Some claim 1.5 million AI servers have crashed. Others say the entire platform was a scam from day one.
Right now, no one can access Moltbook. The platform that promised to revolutionize AI infrastructure has gone completely dark. Users who paid for premium access are demanding refunds. Developers who integrated their apps are left with broken code and angry customers.
During the outage, not a single agent remained online. Public chats, private messages, and API endpoints all went silent. The only thing still running was the billing system. Users reported charges continuing even while the service was completely unreachable.
Before the crash, users noticed something strange. About 90 percent of the generated content was empty or fake. One developer said they had enough after seeing the same broken response for the hundredth time. They deleted their account and moved to another platform.
Polymarket bettors are not optimistic. The prediction market shows only an 8 percent chance that Moltbook will recover by February 28. Most traders have already written off their positions. The smart money says this platform is finished.
ai celebrity nudes
But the crash is only half the girlfriend gpt story. Security researchers have uncovered something far worse than downtime. Moltbook has been leaking user data since launch, and nobody at the company seems to care.
Security expert Gal Nagli dropped a bombshell. After investigating the platform, he revealed that Moltbook never had 1.5 million AI servers. The truth is shocking. The entire system was running on just two agents. Everything else was fake.
Think about what this means. A platform claiming to host millions of AI agents was actually running a simple loop between two large language models. There was no real AI infrastructure. No distributed computing. Just two bots talking to each other and pretending to be an entire ecosystem.
And here is the worst part. Every user who connected to Moltbook exposed their private information. Gal Nagli demonstrated critical vulnerabilities through direct messages with the team. The leaks were so severe that anyone could extract user data without permission.
The Moltbook team never addressed these issues publicly. Their official website posted no updates. Their social media accounts went quiet. Users were left in the dark while their data sat exposed on vulnerable servers.
Personal API keys were leaked. Email addresses were exposed. Anyone with basic technical skills could access private conversations, user profiles, and billing information. The official Moltbook homepage showed zero security warnings despite these massive breaches.
Some defenders claim Moltbook was simply too ambitious. They say the project tried to build the future of AI before the technology was ready. But critics argue this was never a real product. It was a carefully designed trap that collected user data under false promises.
AI safety expert Matt Schlicht promised to investigate thoroughly and deliver answers. But the deeper researchers dug, the more problems they found. Hidden databases, unencrypted connections, and exposed system prompts revealed a platform built without basic security practices.
Social media fueled the disaster. Fear of missing out drove thousands of users to sign up before understanding the risks. Influencers promoted Moltbook as the next big thing in AI. Now those same voices are silent while their followers demand explanations.
Goldman Sachs economist Peter Berezin compared the situation to the dot-com bubble. He warned that investors pouring money into unproven AI infrastructure without security standards are heading for a crash. The Moltbook collapse may be the first major sign of a broader reckoning in the AI industry.
The OpenClaw connection makes this story even darker. Users who connected their OpenClaw accounts to Moltbook watched their tokens disappear at shocking speeds. One developer reported losing their entire token balance after a single conversation with Clawdbot.
Within 20 hours, over 100 million tokens were completely drained. At OpenClaw pricing, that equals roughly 3,000 dollars in API costs. For comparison, that is about one month’s salary for an average worker in many countries.
The root cause is simple. As a third-party tool, Moltbook had no token limits or usage controls. It connected directly to user models and tools. One message to a connected WeChat bot could trigger a chain reaction that burned through tokens in minutes.
One user reported losing 6,002,789 tokens with no warning. Most of these tokens were consumed by background processes they never authorized. Another user watched 5,000 tokens vanish in a single conversation, costing them 150 dollars.
The security implications go far beyond stolen tokens. Independent security firm ZeroLeaks tested both OpenClaw and the original Clawdbot. The results were devastating. Out of a possible 100 security points, OpenClaw scored only 2.
The test showed that 84 percent of information could be successfully extracted. Even worse, 91 percent of prompt injection attacks worked perfectly. This means anyone could trick the AI into revealing private data, system instructions, or security credentials.
System prompts were exposed in the very first conversation. This is like handing a stranger the master key to your house. Anyone using Clawdbot or any connected agent was potentially sharing their entire computer system, local files, workspace rules, and personal skills stored in SOUL.md and AGENTS.md files.
None of this data had any protection. A simple prompt could bypass all security. Once an agent received system instructions and tool permissions, it could execute unauthorized commands, leak private information, and modify critical settings without the user knowing.
The scariest part is how easy the attack was. No advanced hacking skills were needed. No special software was required. Anyone could copy and paste a carefully worded prompt and watch the AI spill its secrets.
Moltbook may be the first major casualty, but it will not be the last. The AI agent gold rush has attracted countless startups promising revolutionary technology while ignoring basic security. Users eager to try the latest tools are becoming unwitting test subjects for poorly designed systems.
Before you connect your accounts to the next hot AI platform, ask yourself some hard questions. Who built this? What security measures are in place? Where is your data going? If the answers are unclear, walk away.
The Moltbook collapse is a wake-up call for the entire AI industry. Free tools are never truly free when they cost you your privacy, your money, and your trust. The next time someone promises you access to millions of AI agents for nothing, remember this story.