It looks like we're back to "move fast and break things" at Meta, with Meta.ai appearing to be off to a troubled start, exposing users' very private search queries in a bizarre public feed. And the company has reportedly been offering $100 million signing bonuses to OpenAI staff?!

News about Mark Zuckerberg's new effort to build a new "superintelligence" AI lab at Meta, separate from the existing division that was already working on AI stuff at the company, was met with celebration by investors last week. But the announcement of Meta's $14.3 billion investment in Scale AI — for a 49% stake in the company — and the recruitment of its founder Alexandr Wang and some of its employees for the new AI team is raising some eyebrows.

OpenAI CEO Sam Altman went on his brother's podcast "Uncapped" today, and spoke about how Meta has been aggressively trying to poach OpenAI's top brass too, offering, he says, signing bonuses in the range of $100 million. And, he says, no one at his company has taken that very considerable bait.

"I respect being aggressive and continuing to try new things," Altman says, with a big caveat. "There’s many things I respect about Meta as a company, but I don’t think they’re a company that’s great at innovation. I think we understand a lot of things they don’t."

Altman added that the strategy of poaching other companies' innovators can backfire too. "That basically never works. You’re always going to where your competitor was, and you don’t build up a culture of learning what it’s like to innovate."

Altman's comments come on the same day as this scathing New York Magazine piece about Meta.ai and it's very weird, initially highly troubling social feed — which was posting users' extremely private prompts and queries, seemingly unbeknownst to them because they didn't see the opt-out button, or something.

This was first reported by Business Insider last month, and some examples include an incarcerated man asking the chatbot legal advice likely pertaining to turning state's evidence, people discussing extra-marital affairs and loveless marriages, and people discussing crimes and tax evasion. All of these queries were being posted in a "Discovery" style social feed for completely incomprehensible reasons — like who at Meta thought this was a good spin on AI chatbot usage?

"The feed is almost entirely boomers who seem to have no idea their conversations with the chatbot are posted publicly," wrote Andreessen Horowitz executive Justine Moore in a June 11 tweet.

The glitchy Meta.ai main-page feed seems to have been cleaned up in the last week, after it was initially just a firehose of this very sensitive content. But the thing still doesn't seem to have a reason to exist, other than to show people what others with weird proclivities — including those who enjoy AI imagery of scary ventriloquist dummies! — are asking the chatbot to do for them.

Still, NY Mag notes, "the public posts are still strange and frequently disturbing. This week, amid bizarre images generated by prompts like 'Image of P Diddy at a young girls birthday party' and '22,000 square foot dream home in Milton, Georgia,' and people testing the new 'Restyle' feature with videos that often contain their faces, you’ll still see posts that stop you in your tracks, like a photo of young child at school, presumably taken by another young child, with the command 'make him cry.'" And all of these are appearing alongside people's actual Meta usernames.

Yes, so, it would seem that this AI chatbot is being used much like a search engine, and likely nobody really wants their searches shared — except those creating meme-like imagery, like one of Donald Trump working as a Walmart cashier.

The future is now, everyone, and it's often sad.

Related: Meta Launches Standalone AI App, Hopes It Will Compete With ChatGPT