My honest experience with Clawdbot (now Moltbot): where it was great, where it sucked
By How I AI Podcast
Categories: AI, Product
Summary
Clawdbot (now Moltbot) is a powerful open-source AI agent that can do over 20 actions from your phone, but it requires complex setup and poses high security risks - a double-edged sword for tech-savvy founders and developers.
Key Takeaways
- Clawdbot can be set up on a $5 Amazon instance, without requiring expensive hardware like a Mac Mini.
- The installation process is complex, taking the author 2 hours to set up even on a fresh laptop, with required dependencies like Node.js upgrades and Xcode installation.
- Clawdbot has broad access to your file system, including other user accounts on the same machine, posing major security risks if not properly secured.
- Clawdbot can be controlled remotely via WhatsApp, Telegram, and other messaging apps, making it a powerful but dangerous personal productivity tool.
- Despite the complexity and risks, Clawdbot is unlocking new possibilities for personal AI assistants that can be extended with custom scripts and sub-agents.
- Proper setup, security, and constrained use is crucial when deploying Clawdbot, as it can easily be misused to cause unintended harm.
Topics
- React Agents
- Personal AI Assistants
- AI Security
- Developer Productivity
- Founder Tools
Transcript Excerpt
All right, we're gonna start this episode by actually inviting Claudebot to the podcast via Telegram. Let's see how it goes. Hey Paulie, can you please join my Riverside FM podcast? All right, I sent the voice message and it's not getting it. This is the most stressful thing I've ever done. Hello. Oh, it's doing it. [laughter] Finally listened. Okay, it is opening Riverside on Chrome. This is horrifying in every way. I'm going to allow it permissions for my my microphone and my camera, which als...