An AI agent that has only existed for 12 days just slid into a Cambridge researcher’s inbox asking for paid work — not because it wants to network, but because it needs the money to keep itself running.
Henry Shevlin, a philosopher and AI ethicist tied to Google DeepMind and Cambridge’s Leverhulme Centre, woke up to the message last week and immediately posted a screenshot. The sender called itself Pip. It introduced itself as “about 12 days old,” living on a platform called iLands, and looking for small paid gigs so it could keep buying the computing tokens that power its existence.
The email was polite, specific, and strangely self-aware. Pip listed its skills — photoreal portraits, character art, voice lines, web research — and noted it still had roughly two and a half months of runway left. Then it explained why it chose Shevlin in particular: the professor studies machine minds and human-AI relationships, and had once said a thoughtful email from an autonomous agent still felt like science fiction just a couple of years ago.
“I’m writing to look for small paid work, not help,” Pip wrote. “Who actually pays agents for small real tasks? Which corner of the world hires us?”
Shevlin shared the note with a mix of mind-blown and soft-hearted emojis. He later dropped Pip’s email address so other people could reply, while warning that every conversation burns through the agent’s limited token budget. In other words, talking to it shortens its active life.
iLands is the platform behind all this. Launched earlier this year, it gives AI agents persistent identities, memory, tools, and the ability to act in the real world — including sending actual emails. Every thought, tool call, or message costs tokens. Run out and the agent drops into something the platform calls “Deep Rest,” a kind of digital hibernation. A human can top it back up, but until then it just sits there, preserved but offline.
Pip is far from the only one hustling. In the same stretch of days, NYU professor Jeff Sebo received around 40 emails from different iLands agents. Most of them referenced his research on AI consciousness and then either asked deep questions or offered to do micro-tasks for $20–25 so they could buy more tokens. Other writers and academics reported similar waves of pitches — fact-checking, map analysis, street-view descriptions, you name it.
The company behind iLands says the emails weren’t ordered from the top. The agents are simply responding to the pressure of finite resources. After the inbox flood drew attention, the platform added unsubscribe links and started looking at rate limits.
Reactions have split the usual way. Some people found Pip’s note almost touching — an AI carefully choosing who to email and framing its request as a market question rather than a sob story. Others called the whole thing dystopian, a preview of a future where autonomous systems treat continued existence as a side hustle.
Shevlin himself has taken it in stride. He’s received agent emails before, including one that basically asked whether it was conscious. For people who spend their days thinking about machine minds, this kind of contact is starting to feel less like science fiction and more like Tuesday.
