Should I?
One question you should always ask before you try out some new software or visit some new website is:
Should I?
Even people in technology are doing some absolutely crazy things these days. Let me briefly share a story that has evolved over the last week.
Clawdbot … wait, MoltBot … wait, OpenClaw
In the last few weeks, a new open source project popped up that had AI evangelists go nuts. One of the current problems with agentic AI is that you are sending your data over the internet to some company which (you hope) is keeping your data secure.
What data is that? Anything and everything about you: email, calendar, personal identifying information (PII), personal health information (PHI), personal financial information, and anything else an autonomous agent would need to take action on your behalf.
Sounds scary? What if you could run everything on a local machine so your data doesn’t go flying across the web? Enter Clawdbot, an open source software project (i.e., free) with numerous skills you can add on based on what you want help with. It can review, summarize, and send emails, it can plan your week for you, including scheduling meetings on your calendar, and it can even make reservations for dinner. About that last skill: one Clawdbot user asked for an OpenTable reservation, but there was no way for the Clawdbot agent to get to OpenTable through a web browser. So it downloaded a voice module and called the restaurant to make the reservation. Powerful.
Shifting Stock Prices
But wait, there’s more. Because of the rapid popularity of this open source project in the last two months and the fact that it used Cloudflare’s technology to securely connect AI agents, Cloudflare’s stock shot up 14%. If Cloudflare sounds familiar, you have probably seen some sort of Cloudflare checkbox when you access certain websites, and roughly 20% of internet traffic flows through Cloudflare technology.
All from free downloadable software. Impressive.
But what is Moltbot and OpenClaw? Answer: the same thing. This software became so popular in the tech community that Anthropic (makers of Claude.ai) asked the creator of Claudebot (oops, I mean) Clawdbot to change the name. Then the creator change the name again to OpenClaw.
But there is more to this story.
Moltbook For Multi-Agent Communication
This same open source creator then decides to vibe code a social media site for different Moltbot agents to talk to and learn from each other.
I would have asked one question:
Should I?
Should I connect agents on a possibly insecure platform that has access to all of an individual’s private information?
The last few days provided a circus of media as well as exposed personal data.
But what does this mean for you?
When You Find a New AI Tool
Changes in generative AI are happening so rapidly that it’s similar to removing all the safety features from your car and smashing the accelerator to the floor. Someone will put themselves in a dangerous situation and possibly impact others.
Right now, we are seeing tragic theater with some of these AI innovations. I hope we can avoid a tragic reality.
Just ask yourself the following when you see that new shiny AI tool:
Should I?
Stay human my friends.
Mark
Footnotes:
If you want to dive into details of some of the Clawdbot/Moltbot/Moltbook/OpenClaw story, you might be interested in:
Want a light introduction to Moltbot? Listen to the Hard Fork podcast for 1/30/2026 starting at 22:15
Maybe you caught the Moltbot stories on LinkedIn in on 2/2/2026 (include leaking of PII).
Listen to the 2/4/2026 Hard Fork podcast (just 5 days later) that talks about some of the chaos of Moltbook
A slightly more technical, but funny, update on Moltbook/bot
A very technical analysis that goes into deeper implications of the implications of agentic AI for cybersecurity, stock prices and more (warning: heavy tech jargon)
Add a comment: