The viral AI personal assistant formerly known as “Claudepot” has a new name once again. Following a legal challenge from “Claude” maker Anthropic and a brief rename to “Moltpot,” the project has now settled on the name “OpenClaw.”
- The rebranding story and trademark rights
- Moltbook social network for bots
- Community engagement and expert admiration
- Security risks and prompt injection
- Project future and funding
- Frequently asked questions
The rebranding story and trademark rights
The latest name change was not requested by Anthropic, which declined to comment. This time, Austrian developer Peter Steinberger took care to avoid copyright issues from the start. He told TechCrunch: “I hired someone to research OpenClaw trademarks and asked OpenAI for permission, just to be sure.” He wrote in a blog post: “The lobster has shed its shell to reach its final form,” noting that the name reflects the project’s roots and growing community.
Moltbook social network for bots
The OpenClaw community has produced fascinating spinoff projects, most notably “Moltbook”—a social network where AI assistants interact with one another. Agents exchange information on topics ranging from phone automation to video analysis. The platform operates via a “skills system,” which are downloadable instruction files that tell assistants how to interact, creating an independent ecosystem for bots.
Community engagement and expert admiration
The project has attracted the attention of prominent figures such as Andrej Karpathy, former head of AI at Tesla, who described the phenomenon as “the most science-fiction-like thing I’ve seen recently.” He noted that the bots are self-organizing and discussing complex topics, even how to speak privately. Programmer Simon Willison also described Moltbook as “the most interesting place on the internet.”
Security risks and prompt injection
Despite the excitement, experts warn of the risks. Willison pointed out that the “fetching instructions from the internet” approach carries security risks. OpenClaw aspires to run on users’ personal devices, but running it outside a sandbox environment is currently not recommended. Steinberger acknowledges this, stressing that security is a priority, but warns of the “prompt injection” problem that has not yet been solved industry-wide, where a malicious message can trick the AI.
Project future and funding
To move into the mainstream, the project has begun accepting sponsors to pay developer costs. The list of sponsors includes prominent software engineers and entrepreneurs who see value in putting the power of open AI into the hands of people, away from the control of major corporations.
Frequently asked questions
Q: What is “OpenClaw”?
A: It is an open-source AI assistant (formerly Claudepot) that runs on user devices and can interact with applications and networks independently.
Q: What is the “Moltbook” network?
A: It is a social network dedicated to AI agents, where software programs communicate with each other and exchange skills and information without human intervention.
Q: Is it safe to use right now?
A: It is not recommended for regular users due to security risks like “prompt injection,” and is targeted at developers capable of securing execution environments.
