The owner of SpaceX, Tesla and Twitter says OpenAI is lying through its teeth about training artificial intelligence.

WCCFTech reported about the New York Times Digital Events Summit where Musk was present. He talked about the current CEO of OpenAI (Sam Altman) and his former head of research (Ilya Sutskever), and said something along the lines of anyone who doesn’t want to advertise on Twitter can go f**k themselves. He also gave some of the reasons for starting OpenAI.

According to him, the company was founded because most of the talent in artificial intelligence was previously at Google, at DeepMind. He criticizes OpenAI for becoming a for-profit company, which he believes is why Altman was temporarily removed as CEO (he has since returned to his post; during his brief stint he turned to Microsoft, but since they invested quite a bit of money in OpenAI, he almost stayed in-house…). Musk has been interested in artificial intelligence since his university days, but he didn’t really get involved because he didn’t know whether the new technology would bring greater risks or benefits. In contrast, he believes that the possibility of life on multiple planets or sustainable energy could lead to positive results right from the start.

OpenAI was born to be open source, but now Musk would rather call it “super-closed-source AI for maximum profit.” “What I care about is the reality of goodness, not the perception of it. And what I see everywhere are people who care about looking good while doing evil. It’s gone from an open source, a foundation, a 501(c)(3), to … all of a sudden it’s like a $90 billion, fully profitable, closed-source corporation. So I don’t know how you get from here to there. But it seems like, I don’t know, how you get there. I don’t know, is this legal? I’m like, is this legal?” Musk said.

He saw Sutskever (whom he brought to OpenAI) as wanting to fire Altman and wanted to investigate why, but Sutskever was unavailable, but ultimately Sutskever’s involvement soured his relationship with Google founder Larry Page. Musk believes that data is more valuable than gold, and that while training on copyrighted data is divisive, OpenAI (and every other company) uses such data to train algorithms to come to their own new conclusions about the technology. Interviewer Andrew Ross Sorkin asked if companies are lying when they say they use non-copyrighted data from AI companies, to which Musk replied, “It’s pure lies, 100%.”

Maybe. Maybe not.

Source: WCCFTech