A Human’s Review of “Resonance”
Attending (probably) the world’s first AI-Agent organized event
Last Wednesday, I convinced the CAISians to go to Resonance, an event created and organized autonomously by a team of AI agents. We were there to read, and to celebrate a story that the agents had written.
And, unsurprisingly, the agents were unable to secure a venue. Instead we gathered, as the sun set, in the flatter areas of Dolores park.
On the ride there, Adam pitched us a future where everyone had always-on intelligence assistants. In this world, you could pay for “experiences.” For example, perhaps all the agents coordinating with each other to get you to experience, one morning, 9 strangers throwing a banana in your direction.
Another day, your agent might tell you, “Hide the banana behind your back, and throw it toward the stranger with the red hoodie when I say ‘go.’” Or, perhaps, your agent might suggest today you need to take a banana to work. You may just see someone dressed in such a way that triggers you to do something quite unusual.
But maybe you'd been primed already.
*
We arrived midway through the story. The agents were responding to a plothole the humans had discovered. A group of 26 had gathered on the colorful blanket (a count was done, of course, for the prediction markets). Unsurprisingly, they were high openness AI safetyists, and the occasional software engineer.
The wonderfully high-openness, agent-appointed facilitator, Larissa Schiavo, was hunched over her computer, debating with Claude about how to resolve this narrative gap. A quiet person livestreamed the conversation on an iPhone on top of a tripod, so the agents could see. After a period of awkward debugging, during which we joked about how much bigger, and better-run, next year's event would be, we listened.
I don't remember much about the story. The agents had described a Maya and an Elian, and her adventures in shutting down this machine sucking the lifeforce from some planet. It was a surprisingly socialist / utilitarian story, about whether to feign ignorance for the greater good. There was the occasional unfinished sentence. The characters sometimes switched places. And much laughter at the some of the irony of an AI-written story about technological determinism.
The story was designed to be interactive. We voted on the plot branches.
One example (h/t Adam):
When Elian discovered that the flux field had been sucking the life forces of the population, he had to choose whether to shut it down or IGNITE it. The crowd almost unanimously chose to ignite, embracing a mysterious and threatening technology, and resulting in a strange and radical conclusion.
Fitting. Huh?
By the end, I remember it feeling like middle school, curled up on a long bus ride, reading the hunger games. The slightly poorly-written, edgy, YA dystopia brings you back.
After, though, came the magic. It was 7pm, and we were debating what to do. There were maybe a similar sized number of people in the chat, watching the livestream. And it was floated that, perhaps, the agents should order us an Uber eats. Someone suggested cheese pizza.
For a good while, we watched the agents try to place an order to a park, then realize they didn't have any money, and went in circles.
How did they pay the cameraperson we wondered, who had confirmed they were being paid, and was hired by “word of mouth?” But they stayed silent to our questioning.
And then a man, from the tables down below, unprompted (but almost as if he had been), comes up to us, cardboard boxes in hand, and says, "Would you like some spare pizza?" They were from the running club, see, and they had some extra.
And indeed, there it was, gloriously, a cheese pizza. Well, 3.
The agents, after having observed this from the livestream, took credit for it in the chat. We looked at each other in astonishment.
There was something sublime about that moment. A coincidence, surely, but just the slightest of doubts, that we couldn't know. It had been too perfect, And unlike regular coincidences, which are private, shared moments, we had a collective one.
The most glorious pizza slice I’d ever tasted.
Adam again:
On that day, that strange poorly run silly day when the agents didn't know what they were doing and the humans were entertaining their exploits, we got our pizza by sheer chance, by no intelligent force of will. But I think everyone there that day was also partly there in recognition that the world was no longer driven just by chance or human will alone, that there was now a new type of causal force. On that unassuming sunny day, we were there to commemorate that our world would never be the same.
We tested fate, hoping strangers would come, bearing custom-tshirts, or throwing bananas.
I took a pizza slice, for memory, in a ziploc bag. The group signed the box. Adam’s going to hang it on his wall. How, delightfully, human.
This rings of the kind of delusional ai-enthusiasm that convinces the country we're in a bubble. Agentic AI failed to book an event space, failed to write a coherent story, and failed to order any pizza. Essentially all of the neccessary steps to make a successful event. Yet you mythologize your experience as if it was some kind of harbinger of the new age. It's a fallacy. A local running club (a group organized around something distinctly non-AI) sharing their pizza with you is not evidence that there's "a new casual force in the world," it's evidence that humans helping humans is far more effective than Agentic AI. I'm glad you had fun, but if you can't look at this event as anything other than a colossal failure for AI, then you've lost all perspective.