Elon Musk, the founder of Tesla and SpaceX, is pushing us towards an increasingly technologically advanced future, one in which we drive around battery-operated cars and take recreational trips to Mars.
But one area of scientific advancement has him deeply troubled: Musk is not so keen on artificial intelligence.
Over the weekend, it appears that the billionaire tech inventor caught a summer blockbuster and did some light reading.
Related: The Robots Are Coming: Zuckerberg, Kutcher Invest in Artificial Intelligence Startup
"Guardians of the Galaxy is great. Best line is when Star-Lord says his spaceship looks like a Jackson Pollock painting under black light," he tweeted on Saturday, before tweeting something a little more serious:
Worth reading Superintelligence by Bostrom. We need to be super careful with AI. Potentially more dangerous than nukes.
Musk is recommending Nick Bostrom's book , which explores a future in which machine learning has surpassed human intelligence. (Spoiler: it's not a pretty picture.)
Related: Does Google's Buying Spree Mean the Robot Apocalypse Is Near?
Clearly, the book had a deep impact: Musk followed up the aforementioned tweet with this:
Hope we're not just the biological boot loader for digital superintelligence. Unfortunately, that is increasingly probable
This isn't the first time Musk has expressed deep reservations about the future of A.I. He is an investor in DeepMind and Vicarious, two A.I. ventures, but “it’s not from the standpoint of actually trying to make any investment return," he told CNBC in June. "I like to just keep an eye on what’s going on.”
Because in Musk's view, 'what's going on' could potentially result in a 'Terminator'-like future. “Nobody expects the Spanish Inquisition,” Musk said. “But you have to be careful.”
Unfortunately, space travel isn't going to save us. If robots start to turn on us, couldn't we "escape to Mars if there is no other option?" CNBC’s Kelly Evans asked, half-joking.
But Musk took the question seriously: “The AI will chase us there pretty quickly,” he responded.
Related: Intel Wants to Make Computers Think More Like Humans