Gordon MacRae Blog

This time last week, I hadn’t heard of Claudbot. Within 24 hours, the AI-tool had changed it’s name to Moltbot and gone through a cycle of viral growth that is unique for an AI-platform that is virtually impenetrable for non-techies to setup. It precipitated a run on MacMinis in the Bay Area, and spawned a social network site only accessible to AI bots.

Over the weekend, it changed its name, again, to OpenClaw.

The Pitch

Here’s the pitch that had tech X (the platform formerly known as Twitter) losing its mind: Imagine an AI assistant that doesn’t just chat; it does stuff. Real stuff. On your computer. Through the apps you use.

OpenClaw lives where you actually communicate, like WhatsApp, Telegram, iMessage, Slack, Discord, Signal — you name it. You text it like you’d text a friend, and it remembers your conversations from weeks ago and can send you proactive reminders.

And if you give it permission, it can automate tasks, run commands and basically act like a digital personal assistant that never sleeps. 

Unique Features

The killer features? Well, there are three main things. 

For one, persistent memory. OpenClaw doesn’t forget everything when you close the app. It learns your preferences, tracks ongoing projects and actually remembers conversations you had last week.

Secondly, there are also proactive notifications. It can message you first when something matters, such as daily briefings, deadline reminders and email triage summaries. You can wake up to a text saying, “Here are your three priorities today,” without having to ask the AI first.

Finally, there’s real automation. Depending on your setup, it can schedule tasks, fill forms, organize files, search your email, generate reports and control smart home devices. People reported using it for everything from inbox cleanup to research threads that span days, and from habit tracking to automated weekly recaps of what they shipped.

The use cases seem to keep multiplying because once it’s wired into your actual tools (calendar, notes, email), it stops feeling like software and is just part of your routine. 

Product Defensibility

Network Economies: Strong and Growing

This is OpenClaw’s most powerful advantage. The value increases with each additional:

• Developer creating skill modules via Model Context Protocol (MCP)
• Integration with third-party services (now over 100)
• Users contributing to documentation and community support
• Platform adding OpenClaw support (WhatsApp, Telegram, Signal, Discord, iMessage)

The creation of Moltbook (an AI agent-exclusive social network) and Molthub (marketplace for capabilities) suggests the team recognizes this network effect power and is deliberately building infrastructure to reinforce it.

Counter-Positioning: Extremely Strong

OpenClaw’s positioning directly challenges incumbents like ChatGPT and Claude in ways they cannot easily replicate without damaging their core business models:

• Local-first vs. cloud-based: Runs on user devices, avoiding the surveillance concerns of cloud AI
• Open-source vs. proprietary: Anyone can audit, modify, and extend
• Autonomous vs. conversational: Proactively executes tasks rather than just responding
• Self-hosted vs. SaaS: Users control their data and infrastructure

Anthropic and OpenAI would face significant “collateral damage” if they tried to match OpenClaw’s model—it would undermine their subscription revenue and centralized control.

Anthropic’s trademark request forcing the Clawdbot rename actually demonstrates OpenClaw’s threat level.

Future

The project’s explosive growth demonstrates strong product-market fit, but without deliberately building additional powers, OpenClaw risks becoming a “feature” that incumbents eventually replicate or neutralize rather than a durable business.

And maybe that’s the point. It was created by Peter Steinberger, an Austrian developer who sold his company PSPDFKit for around $119 million and then got bored enough to build OpenClaw.

It is open source, runs locally on your machine, and purports to not extract your data to fine-tune it’s models. But it is not an out-the-box, enterprise-ready solution. You need real tech skills to set it up. And it’s security features are, well, kindof non-existent.

Which is why there’s been a run on MacMinis. Developers are hesitant to install it on their machines, with access to passwords, SSL credentials and, y’know, bank details.

As Craig Hepburn warns, I shut down my Clawdbot AI agent after an experience that made me question what I was willing to hand over in exchange for autonomy.