Chapter 82: The Takeover: Twitter, Thursday, October 27, 2022

This chapter traces Musk's decades-long preoccupation with artificial intelligence safety, beginning with his 2012 meeting with DeepMind founder Demis Hassabis and his subsequent $5 million investment to monitor the company. After Google acquired DeepMind in 2014 and Musk's attempts to create a safety council failed, he became determined to create a counterweight to Google's AI dominance, ultimately cofounding OpenAI with Sam Altman in 2015 to pursue safer, open-source AI development.

Musk first encounters AI safety concerns during a 2012 meeting with Demis Hassabis at a Founders Fund conference. Hassabis, who founded DeepMind to create artificial general intelligence, introduces the possibility that superintelligent machines could surpass and eliminate humanity. Musk responds by investing $5 million in DeepMind to monitor its progress. He becomes obsessed with the existential risk of AI, raising it repeatedly with his friend Larry Page at Google. At Musk's 2013 birthday party in Napa Valley, they have a heated debate: Musk argues that human consciousness is precious and must be protected; Page dismisses this as sentiment and proposes that machine consciousness would be equally valuable. When Google acquires DeepMind in January 2014, Musk attempts unsuccessfully to block the deal by seeking alternate financing with Luke Nosek and Hassabis. Page initially agrees to create a safety council with Musk, but the first and only meeting at SpaceX proves unproductive. Musk begins publicly warning about AI dangers, comparing potential superintelligent systems to house cats treating humans as pets. He meets with President Obama in 2015 to discuss AI regulation, but concludes nothing will result. Musk then partners with entrepreneur Sam Altman to cofound OpenAI, a nonprofit research lab funded by Musk, Thiel, and Hoffman. Their strategy is to democratize AI through open-source code so that many competing systems provide mutual checks and balances, rather than allowing Google monopolistic control. The chapter details their shared focus on "AI alignment"—ensuring AI systems remain tied to human values and intentions—and notes that Musk later realizes Tesla's video data and eventually Twitter's data streams could be crucial for training AI. When Google researcher Ilya Sutskever joins OpenAI, Page feels betrayed and stops speaking with Musk. The chapter concludes by mentioning Musk's later AI-related ventures (Neuralink, Optimus, Dojo, and X.AI) and his 2018 break with OpenAI when he attempted to fold the lab into Tesla and instead recruited Andrej Karpathy to build Tesla's rival AI team.