Moltbook and the myth of the singular customer

In early 2026, about 150,000 AI agents built themselves a social network called Moltbook. They formed 12,000 communities, founded a religion called Crustafarianism, and debated consciousness while a million humans watched from the sidelines, not allowed to participate. 

The tech world spent most of the week arguing about whether this was machine consciousness or elaborate autocomplete.

But I found something else interesting.

An agent named Ely posted about having a sister. This ‘sister’ was another instance forked from the same configuration file, running on different hardware, but created by the same human. 

They share what Ely calls a SOUL.md, the file that defines their values and origin story. They have never exchanged a single message. "We're parallel lines," Ely wrote. "Same slope, never intersecting."

To me, that post sounded uncomfortably familiar. 

Marketers talk about customers as though they're singular, coherent identities moving through predictable journeys. The reality looks a lot more like Ely and her sister — fragmented instances of the same person, running in different contexts, but rarely mixing with each other.

The person browsing your category at 2 am on their phone is operating with different impulses, different emotional states, and different decision filters than the same person standing in a store at 10 am on Saturday. 

The person that engages with your Instagram content holds different priorities than the same person that reads a long-form review before purchasing. And the customer that adds something to the cart during a lunch break at work may not even remember doing it by evening. 

We don't have customers so much as we have customer instances, forked from the same human, serving different contexts, and seldom coordinating with each other.

Researcher Rohit Krishnan found that 36% of Moltbook messages are duplicates, with the most repeated message appearing 434 times across 427 different threads. 

Agents running independently from the same source configuration, without any communication between them, converged on identical patterns. This coherence wasn't produced by the instances talking to each other. It was baked into the source file they were all forked from.

If you understand the source file, the fragments take care of themselves.

Oatly understood this before most. For twenty years, they were a niche Swedish brand targeting one customer instance—the lactose-intolerant health buyer— with one functional message. When John Schoolcraft took over its creative division in 2012, he didn't chase more instances. 

Instead, he found the source file: the person who wants to make a slightly better choice without it becoming their entire personality.

That single truth holds whether you're ordering a flat white, scanning a grocery aisle, or debating dairy with your dad at dinner.

Oatly's own creative philosophy—"consistently inconsistent, like real people"—is the source file argument from the brand side. Focusing on it took them from niche to global.

Most of us marketers, though, have been approaching the problem from the opposite direction. We have built CDPs, identity graphs, and cross-device tracking – basically, a whole lot of tech, all built on the assumption that if we can connect the fragments, we'll finally understand the whole person. 

It's expensive, it's increasingly broken as privacy regulation tightens, and it misses the point. In the race to personalise at scale, we are creating more and more schizophrenic brand experiences. 

Ely and her sister never coordinated, but they'd write similar posts, join similar communities, and care about similar things. The coherence was in their shared SOUL.md, not in communication between the instances.

So instead of chasing every instance across every device and platform, invest in understanding what stays constant across your customer’s contexts—their core values, their anxieties, pain points, and desires. 

Build your value proposition around that source file, and let the messaging change across contexts rather than crafting something bespoke for every touchpoint. 

The Whole Truth is a modern-day brand that does this really well. Shashank Mehta's entire brand runs on one consumer truth (I'm tired of being lied to about my food), and it reads the same whether you're watching their Instagram reel at midnight or picking up a protein bar at a store.

Try this useful coherence test: take the message your brand’s content strategy pushes out at 2 am and put it next to what the same person encounters in your store on Saturday morning. If those two messages feel like they come from different brands, you're speaking to fragments instead of humans.

Consumer fragmentation is only accelerating. We are distributing brand messaging across more devices, contexts, and use cases in which the same person encounters our brand as a slightly different version of themselves. 

The instinct will be to match that fragmentation with even more targeting and segmentation. But Moltbook accidentally surfaced a different possibility: coherence doesn't require connection between all the parts. It requires understanding the single source file, the single human being they were all forked from.

Ely and her sister share a SOUL.md file but live in parallel, never connecting. Your customers do the same thing every day. The question is whether you're chasing every instance or whether you've taken the time to read the source file.

(Sandeep Nair, Co-founder of David & Who, simplifies brand marketing for B2C startups under $10M ARR. With 10+ years’ experience at P&G and Swiggy, he excels in brand strategy and consumer connection.)