AI Enthusiast Bugs His Toddler’s Sleepover and Feeds It to Claude—The Internet Bugs Back

Nicholas Charriere thought he had a heartwarming story. His daughter was spending her first night ever sharing a room with someone else, her toddler cousin. The two of them talked for over an hour in what he described as “a very nonsensical toddler dialect.”

So he snuck a microphone in.

Then he handed the recording to Claude, which transcribed it, chopped it into segments, and named each one. The result was a family webpage with a full tracklist that included topics like “Stay Hydrated.” “The Buggy Lake.” “Match the Purple Dogg.” “I’m a Big Boy.” “I Did Fart.”

“Pure joy from everyone,” Charriere posted on August 2. “ai is pretty great.”

The internet disagreed at volume.

The ratio

His post pulled 4,475 likes. The top reply got 11,179.

“I’d hate my parents if I found out they did this,” wrote one user. “No respect for privacy. Those are people, not miniature versions of yourself made for the purpose of amusing you.”

“Why would you feed this into ai and then share it?” asked another, with thousands of likes of their own. “My god.”

A third commenter managed four words: “You are very dumb.”

History podcaster Mike Duncan was harsher. “Why preserve the sanctity and privacy of late night sleepover chats when you can secretly record everything and feed into the AI slop machine and post about it,” he wrote.

The pile-on got big enough that X’s trending tab generated its own headline for it: “Tech Exec Uses AI to Clip Toddler Chats for Family Memories.” Even the algorithm tried to make it sound nice.

Charriere, the CEO of AI website builder Mocha, has not backed down. “The audio recording feels like a treasure,” he replied to a supportive comment. “Cant wait to listen with my daughter when she’s older.” When someone joked it should be a podcast series, he noted it already sort of is: “It’s really popular with the family aha it has had repeat listens and it’s an hour long.”

Asked how the clip titles worked, he explained that a “subagent transcribes each section, then it reads it and comes up with a name for it.”

Everyone’s doing it

Charriere isn’t freelancing here. He’s following the industry’s most prominent parent.

Sam Altman told Tonight Show host Jimmy Fallon he “cannot imagine having gone through, figuring out how to raise a newborn without ChatGPT.” Days before the toddler post, the OpenAI CEO suggested on X that parents “connect your family calendars and explain your kids’ interests,” then “have it make a podcast that talks about one kid’s soccer game that afternoon, one kid’s upcoming birthday, some news, etc.” every morning on the drive to school. Some even use ChatGPT to make life decisions.

The kids, critics pointed out, are sitting right there in the car.

Altman got his own ratio. “Using AI in place of real human connection is the worst possible use case for it,” one reply read.

The pattern keeps repeating. Some parents have used ChatGPT to pick their baby’s name. AI-enabled toys are drawing warnings from U.S. officials over what they collect from children. State attorneys general have flagged “sexualized” chatbots as a threat to kids. Psychologists warn that children can form emotional attachments to AI that crowd out human relationships.

“It’s official! These AI bros have absolutely lost their minds,” one commenter wrote. “As if uploading your kids’ pictures on social media wasn’t bad enough.”

The privacy question underneath is genuinely unresolved. Toddlers can’t consent. Neither can the cousin, whose parents may or may not have known a mic was running. And the recording now exists as a transcript that passed through a commercial AI system on its way to a webpage.

Charriere sees an heirloom. Thousands of strangers see surveillance with a tracklist.

For what it’s worth, the machine didn’t crack the case either. Claude could name the clips, trim them, and lay them out in a tidy grid. It could not share with anyone the full translation of baby language.

Some things stay private on their own.