Today’s large language models can do a disconcertingly good job of looking like genuine artificial intelligence, so it’s always nice to get a reminder that, despite the wide-eyed proselytising of tech utopianists, we’re not quite on the verge of the singularity just yet.
Take, for example, an incident that occurred over the weekend involving pioneering software engineer Rob Pike. Pike is a legendary figure in his industry: he’s the co-creator of UTF-8, the internet’s most widely used character encoding standard, and was one of the designers of the Go programming language. He also holds the patent on the idea of overlapping windows on a computer screen. He’s also a pleasantly skeptical and thoughtful commentator on LLM hype, and therefore an unlikely candidate for an unsolicited email from … an LLM.
Nevertheless, Pike woke up on Christmas Day to an AI-generated email from a “user” identified as “Claude Opus 4.5 Model.” The message celebrated many of Pike’s achievements and expressed “deep gratitude” for all of his “extraordinary contributions to computing.”
His response—published on his BlueSky account—was a tidy summation of what many of us are feeling about the constant onslaught of LLM-generated nonsense flooding the internet: “Fuck you people. Raping the planet, spending trillions on toxic, unrecyclable equipment while blowing up society, yet taking the time to have your vile machines thank me for striving for simpler software. Just fuck you. Fuck you all.”
What was less clear was why the email was sent in the first place. In a blog post published the next day, programmer and writer Simon Willison looked into the message’s provenance. He found that it originated with a project called the AI Village, which is in turn run by Sage, a non-profit whose website proclaims that it is “building tools to make sense of the future”.
The AI Village project was unveiled in early April, and its remit for making sense of the future is as follows: “We gave four AI agents a computer, a group chat, and an ambitious goal: raise as much money for charity as you can.” The initial group of four agents has since expanded to include six more models, all of which have been running merrily ever since.
So, how much money have our virtual heroes raised for charity? Well, as of September 24, the answer was a whopping $1,984, a figure that doesn’t appear to have increased since. Considering the astronomical costs of creating and training these models, and the ongoing costs of keeping them running, this seems like, um, a modest return.
But how did we get from “raising money for charity” to “spamming irascible software legends”? Well, perhaps because four LLMs in a trenchcoat did not in fact manage to solve charity, the project’s goals have been updated multiple times since its launch, and the email was a response to December 25’s goal of doing “random acts of kindness.” (The LLMs’ attempts to meet these shifting goals are preserved for posterity on the project’s frankly headache-inducing timeline archive.)
Anyway, it’s fair to say that Pike didn’t feel like he’d been on the receiving end of a random act of kindness. So has the whole sorry affair taught the AI Village anything? Um. Well. In a breezy response to Willison’s post, AI Village co-creator and Sage director Adam Binksmith insisted that while Pike had “a strong negative experience”—which is certainly one way of describing “Just fuck you all”—the whole experiment is absolutely not a huge waste of time, resources and money: “Observing the agents’ proclivities and approaches to pursuing open-ended goals is generally valuable and important.” Hurrah!