OpenAI’s latest “innovation” by sliding ChatGPT into our group chats feels suspiciously like déjà vu for anyone who remembers the internet before it went glossy. The pitch is shiny: up to 20 humans and one omnipresent AI coexisting in a single conversation, coordinating trips, debating ideas, maybe planning who brings what to Tahoe. But underneath the veneer? It’s Usenet. Again. Only this time, the AI reads everything.
Back in the pre-web days, Usenet was the messy, glorious proto–social network where humanity came together to argue, pontificate, overshare, flame, troll, apologize, and repeat. Each newsgroup was a micro-civilization, alt.sports.hockey, rec.arts.books, comp.sys.mac, all powered by shared interest and a total lack of adult supervision. Everyone posted. Everyone watched. And everything lived forever.
Now replace the nerds with 800 million ChatGPT users and substitute the flame wars with “collaborative productivity,” and voilà: GroupGPTNet. Tap an icon, drop in 20 people, and suddenly your weekend-planning thread looks like alt.rec.travel with a bot that never sleeps. At least Usenet didn’t read your messages to improve its neural architecture.
OpenAI says this is about “coordination” and “conversation.” Sure. And Usenet was about “academic file sharing.” Let’s not kid ourselves: the real story is data. Group dynamics are gold mines for model training, the unwritten social signals, the conversational arcs, the norms of humor, tension, conflict, and reconciliation. Companies don’t need to reverse-engineer human interaction when we willingly stage it for them.
And the competition? Even more Usenet-esque. Continua’s bot wants to join your existing group chats, the digital equivalent of injecting an AI into your favorite alt.* channel and hoping no one notices it’s not sentient. It claims less friction. Translation: fewer people have to move, so more conversations get slurped into the machine.
Will people actually talk to bots in groups? Today, it still feels like asking Alexa a question in a crowded bar, everyone freezes, someone snickers, and you regret everything. But social behaviors are malleable. If we normalized AOL chat rooms, emoji as punctuation, and Instagram therapy, we’ll normalize bot-lurkers too.
In the end, OpenAI hasn’t broken ground. It’s broken the seal on nostalgia. Group chats with AI are just Usenet with better UI, worse boundaries, and a business model built on harvesting the digital anthropology of our lives. This isn’t the future — it’s the past, reissued in high resolution.