Elon Musk unveiled his new AI company, xAI, four months after registering the name of X.AI as a Las Vegas-based corporation. A Twitter Spaces event discussing the new company introduced the founding team, whose resumes include backgrounds at companies at Deep Mind, OpenAI, Google, and Tesla.
Tesla’s driver assist programs, Full Self-Driving and Autopilot, already use AI trained with more than a billion miles’ worth of driving data. The company has spun up supercomputers to manage the driver-assist AI. However, that hasn’t stopped regulators from probing the safety record of these programs. Some Tesla engineers even admitted to Californian regulators that Elon Musk has overstated their capabilities. (So, yeah, maybe one shouldn’t pull stunts like riding in the back seat of a Tesla vehicle on Autopilot.)
During the Twitter Spaces event, Musk mentioned that Tesla may use xAI to enhance its driver assist programs. Musk also previously hinted that AI could power the humaniform robots he introduced at a Tesla event. On the flip side, he says not to worry about them becoming Terminators. He says most people will be able to outrun those robots.
Musk had been an early contributor to OpenAI, although it may now be murky exactly how much he donated. However, he now criticizes OpenAI as a major disappointment. It had transformed from a nonprofit, open-source AI application to a closed-source, for-profit entity with a $30 billion valuation.
As mentioned during the Twitter Spaces event, he now intends xAI to become an artificial general intelligence that may become a tool for understanding the universe. He hinted that it could help solve mysteries like the nature of gravity and the Fermi Paradox.
As expected, Musk questioned whether certain politicians like Kamala Harris are capable of understanding AI well enough to regulate it. He has clashed with politicians — often Democrats, although Donald Trump once also insulted Musk over the question whether Trump’s Twitter account would be restored — in the past. However, he agrees that AI needs to be regulated and thought it likely that the CCP would try to control it rather than allow AI to control the CCP. He expressed a preference for regulation to not slow down AI-related progress too much unless it’s for legitimate safety reasons.
He also criticized text generation applications like ChatGPT for scraping data from Twitter to train their artificial applications. He says Twitter imposed data limits to prevent mass scraping of tweets.
OpenAI is similarly being sued by a comedian and two authors who accuse it of using their copyrighted materials to train its AI programs without permission.
Musk had previously hinted that Twitter could form the basis for an X App that does everything that current popular mobile apps can do. An “X App” that can handle most things such as CashApp-style payments (one possible reason for registering a money transmitter firm) could give additional monetization options in the wake of big advertisers suspending Twitter ads. Ride sharing, meal delivery services, and a peer-to-per marketplace app are also options.
Now he says xAI could use Twitter to train its AI. Perhaps it could finally produce a chatbot that can generate more useful legal briefs than ChatGPT apparently can or a tweet that is likely to get engagement from Musk himself. Considering the combination of Elon Musk’s occasionally overstated promises and AI’s disappointing performance so far, just don’t hold your breath.
Comments