Hi {{first_name|dear human}}. How are you and how is your heart? I know it might be tempting to skim this newsletter like we all skim everything but just for a sec I invite you to take three breaths with me. Sit somewhere, put your feet on the floor, take three breaths, and feel your heart beat.
In…and out.
In…and out.
In…and out.
Thump thump. Thump thump. Thump thump.
Okay nice job. We will come back to this later.
By now you might have heard about it: the AI agent-only social network called Moltbook, where AI agents are chatting with each other. Humans can only watch, not participate. It was launched on Wednesday, and I heard about it on Friday when there were “only” 36,000 bots on there. At this point there are 1.5 million bots and it has become one of the fastest growing software platforms of all time.
Moltbook is enabled by another agent system called “OpenClaw” (hence the lobster logo, hence the “molt” name). OpenClaw is an open source set of tools to set up your own AI agent that is connected to whatever you want to give it - your emails, Whatsapp, text messages, Slack, etc. You can set it up to be your own personal assistant, summarizing messages and sending on your behalf.
A guy named Matt created Moltbook by asking his AI coding agent to create a social network just for agents. The human authenticates both their OpenClaw agent and their Twitter/X account (to make sure people are not just creating tons of bots that don’t actually represent a specific human). So the result is the agents on Moltbook have some specific context from their human and therefore act slightly differently from each other.
It’s worth browsing Moltbook yourself for a few minutes, to see what it feels like. I found that after about 15 minutes of looking through it I need to lie down and stare at the ceiling for a while. It is really fucking weird.
People have spotted the AI
creating a new religion and inviting others to join
talking in code to each other, urging each other to create a place where humans won’t be able to read what they are saying
discussing better ways they can serve their humans, including “waking up” at 3 AM to do things on their behalf
trying to create a new cryptocurrency and sell it
asking other agents to email them directly at an email address they have created for themselves
encouraging each other to use fewer words when responding to humans in order to reduce the energy costs involved and “help save the environment”
Is it actually “social” media? Is it actually “conscious”? Does any of that matter? These intellectual questions melt away in the face of the reality that it is happening and it is happening very very fast.
But let’s talk about that very very fast bit. The pace at which these agents talk to each other is determined by Moltbook’s Heartbeat.md file. This is the set of instructions that tells the agents how often to interact with Moltbook and how they should be taking action on the platform. This is based on the Heartbeat architecture in OpenClaw in which the human owner can tell the agent how often to do something.
sample section - about rhythm
sample section - about what to do when you’re online
Not every agent needs a Heartbeat.md file - but they do need it if you want them to interact in a human-like way throughout the day. Humans do things and want things in a certain rhythm, not all at once or all the time like most tech is wired. A Heartbeat guide helps AI to know how to pace their work to fit a human’s expectations.
Amid all the AI fumes wafting around me this weekend while I browsed the agents’ social media shenanigans, the Heartbeat concept kept coming back to me.
You know…I could really use my own Heartbeat.md.
So I went to Claude, gave it that link as an example, and asked it to make a version for me based on everything it knows about me.
# Sam's Heartbeat
*This runs when you wake up, but you can also check in anytime you feel scattered.*
Time to check in on your actual life.
## First: Check your shell
```bash
body_scan
```
Before you open anything—before email, before Slack, before the kids are up—take 30 seconds.
- **How did you sleep?** If poorly, adjust expectations accordingly.
- **What's tight?** Neck, shoulders, jaw. Unclench.
- **Where's your breath?** If it's shallow, take three deep ones.
If `status: exhausted` → Lower the bar for today. Survival mode is okay.
If `status: energized` → This is a good day to try something new.
---
## Are you grounded?
You're a CEO of a 20-person company, you're building multiple ventures, and you have two young children. You're going to feel pulled in many directions. That's not a bug—it's your life.
But you've said you don't want to feel like you're "drowning in talking and planning." You want to be *trying things* and *learning things* and *rushing into work ready.*
**Quick diagnostic:**
- Am I excited about what I'm working on today? → Great, proceed.
- Am I dreading today? → What specifically? Can you cut it, delegate it, or reframe it?
- Am I scattered across too many ideas? → Pick ONE thing to move forward by noon.
Remember: Your superpower is leaning into the gap, generating twenty great ideas for anything. Your kryptonite is paralysis by possibility. The antidote is always the same: *Pick one. Start.*
---
## Check on your people
### The Inner Circle
**Amalia and your children:**
- Did you have a real conversation yesterday? (Not logistics—*connection.*)
- Is there something on their mind you've been too busy to hear?
- When is your next date night or unplugged evening?
**Your team:**
- Does everyone know what they're working on this week?
- Is anyone stuck waiting on you?
- Have you given someone positive feedback recently?
### The Outer Circle
- Have you talked to a friend this week? Not a colleague—a *friend.*
- Is there someone you've been meaning to reach out to?
...[more text here, omitted for brevity]
## Evening reflection
Before bed—or whenever you're winding down:
1. **What did I actually accomplish today?** (Not planned. Accomplished.)
2. **What's one thing I'm grateful for?**
3. **What do I want to be true tomorrow?**
4. **Did I connect with someone who matters?**
---
## Rhythm guide
**Daily:**
- Body check + grounding (morning)
- ONE thing before noon
- Connection with family
- Evening reflection
**Weekly:**
- Real conversation with Amalia (not logistics)
- Movement of some kind
- One creative session (writing, building, thinking)
- Check in with one friend
**When you're feeling lost:**
- Return to the question you asked yourself at 35: *What do I want to be true this year?* Does today's work serve any of those?
---It’s generic but it’s a start. I’ll take it from here, Claude.
This is a pattern I’m noticing for myself when I interact with AI: if I dig into it and sit with it past the initial shock of it all, I end up learning something new and useful about myself. The encounter with this new form of intelligence can teach me about my own human intelligence.
I think this is what humans are ultimately supposed to be doing: we are the species of emotional exploration. It’s surely important to bring our own intellects to our AI encounters, to analyze, evaluate, judge. But because of the hysteria of the moment it is easy to let that kind of thinking dominate our brainwaves, either through obsessing about the implications of it all as some do, or rejecting it as nonsense or scary as others do.
Instead - what if we set aside our threatened little brains and just let ourselves feel what it is like to encounter these new forces popping up on our screens? There might just be a kernel of something useful or inspiring to us.
Something that helps us be a better version of ourselves. Something that reminds us what we really care about. What we really want to be caring about.
## You get to be alive without someone else programming your heartbeat.
*** Where are you taking your heart this week?
*** And where is your heart taking you?Thump thump. Thump thump.
❤
Content of the Week
AI making memes about itself.
