The whole experience reminded our senior AI editor, Will Douglas Heaven, of something much less interesting: Pokémon.
In 2014, someone made a Pokémon game where anyone could control the main character online via the streaming platform Twitch. It was as difficult as it sounds to play, but it was incredibly popular: at one point, a million people were playing the game at the same time.
“It was another strange online social experiment that was picked up by the mainstream media: What does this mean for the future?” Will says. “Not much, as it turns out.”
The craze over Moltbook had a similar vibe to Will’s, and it turns out one of the sources he spoke to was thinking about Pokémon as well. Jason Schlozer, of Georgetown’s Sarros Center for Financial Markets and Policy, saw the whole thing as a kind of Pokemon battle for AI enthusiasts, creating AI agents and deploying them to interact with other agents. In this light, the news that many AI agents have been instructed by people to say certain things that make them appear sentient or intelligent seems more plausible.
“It’s basically a spectator sport, but for language models,” Lowell said.
Will wrote an excellent article on why Moltbook doesn’t provide a glimpse into the future like it has been claimed to. He points out that even if you’re excited about the future of agentic AI, there are some key pieces that Multbock points out are still missing. It was a forum for chaos, but a truly useful hive mind would require more coordination, shared goals, and shared memory.
“More than anything, I think Moltbook has been the joy of the Internet,” Will says. “The biggest question that leaves me now is: How far will people push AI just for a laugh?”
