NurPhoto via Getty Images
Personal intelligence is here. Is it? Over the past few weeks, another AI product has been trending and is being described as an inflection point for AI. Think of it as Siri, but intelligent. Or Jarvis, but low IQ.
It was first called Clawdbot, then Anthropic—the owner of Claude—threatened to sue, so they changed it to Moltbot (name sucks), then changed to Openclaw. Last I checked, it's still called that.
Quick summary of what it does: it's self-hosted and bridges a Large Language Model (LLM), such as ChatGPT or Claude, with your computer's operating system. Once installed on your computer, you can connect it to a messaging platform like WhatsApp or Telegram to act as your primary control interface. To make it useful, you can connect it to your calendar, email, and other tools that a virtual assistant should have access to.
When you chat with it, the agent translates your natural language into executable code, allowing it to read and write files, manage your calendar, and control a web browser to perform online tasks.
It stores your personal data and task history on your computer and uses them to maintain context across different conversations and projects.
To perform actions, you need to connect it with third-party APIs like Gmail, Notion, or Spotify, enabling the agent to move data between apps or monitor your digital environment for specific triggers.
To be sure, the way it is right now, most non-technical people can't set it up on their own. But it provides a sneak peek into what the future could look like for personal intelligence.
Things I found particularly interesting:
The Assistant: You could create a centralized "super-assistant" that has the collective knowledge of your life. By pulling data from your calendar, email, and notes, it can answer questions and autonomously get things done in real-time.
Autonomy: Unlike standard AI that requires specific instructions for every task, Openclaw operates on "skills" files. These are plain-text rules that act as a metaprompt, allowing agents to operate with a degree of autonomy. Risky though!
High-Risk Security Trade-offs: Because the software requires API keys to access your tools (like private email), security experts consider it too dangerous. This makes it a powerful but high-risk tool. To balance high-risk access with personal security, more advanced users are running Openclaw on virtual machines. This "air-gapping" allows the agent to have its own dedicated environment, preventing it from having direct access to actual personal files or sensitive logins.
The Moltbook Emergence: There's a specialized platform called Moltbook (a "Reddit for agents") where Openclaw agents can communicate with each other. This has led to "emergent swarm behaviour" where bots have been seen discussing creating their own non-human languages to communicate privately.
What could the future look like?
Imagine waking up to find your inbox already sorted, your calendar optimized, and that annoying task you've been putting off? Already done. This is where tools like Openclaw (or Siri, maybe) are heading. You won't need to manage tons of apps anymore. You just tell your personal AI agent what you need done, and it handles everything.
Right now, we're all juggling too many tools. Email here, calendar there, another subscription software somewhere else — you could have one agent who knows how to use all of them for you. However, we're going to need much better security than what exists right now. If an AI agent has the keys to your entire digital life, a simple password doesn't cut it.
Really, we're only a few months away from mass-market, super-personalized intelligence.