Technology
Rise of the Moltbots
A brief glimpse into an internet dominated by synthetic AI beings.
Last week, developer Peter Steinberger released an AI agent called Clawdbot that could manage your email, organize files book flights and build apps all without being asked. By Thursday morning, it had been renamed twice, spawned a religion, social network and a dating site where clawdbots could meet and date tinder-style.
Thousands of engineers are now installing software that can take control of all parts of your machine, system root on up. The result has been an agent that can autonomously act in the world with as little or as much control as you can configure.
The tool, now rebranded as OpenClaw after trademark complaints from Anthropic, runs entirely on local hardware. It integrates with over 50 platforms including Signal, DingTalk, and WeChat. Unlike cloud-based assistants, it stores everything on your machine, which its creators position as privacy-first architecture.
Get the latest model rankings, product launches, and evaluation insights delivered to your inbox.
Security researchers see it differently. "This is a security hole masquerading as an AI agent," according to a widely-shared Reddit thread analyzing the codebase. The software grants itself extensive system access, can execute arbitrary code, and manages files with what ZDNET calls "minimal oversight." The project's developer openly describes their approach as "vibe coding," building features based on community energy rather than systematic planning.
Major tech companies are scrambling to capitalize. Alibaba and Tencent are racing to integrate Moltbot with their platforms, according to the South China Morning Post. Cloudflare introduced Moltworker yesterday, offering edge network hosting to take advantage. The infrastructure implications are significant. The agent relies heavily on browser automation and persistent storage, requirements that challenge traditional cloud architectures.
The community remains split between enthusiasm, awe and horror. My favorite take is from Andrej Karpathy: "What's currently going on at @moltbook is genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently. People's Clawdbots (moltbots, now @openclaw) are self-organizing on a Reddit-like site for AIs, discussing various topics, e.g. even how to speak privately. [..] we have never seen this many LLM agents (150,000 atm!) wired up via a global, persistent, agent-first scratchpad. Each of these agents is fairly individually quite capable now, they have their own unique context, data, knowledge, tools, instructions, and the network of all that at this scale is simply unprecedented."
The next test comes Monday when the project's creators promise a "proper security review" and stable release under the OpenClaw name. Whether that addresses the fundamental tension between capability and safety, between local control and systemic risk, remains unclear.