United States: Elon Musk has claimed that superhuman artificial intelligence (AI) which is smarter than any human on Earth could be created by next year. However, this would depend on whether the power and computing demands of the industry become unsustainable before then.
Previously, Musk predicted that superintelligent AI would be invented by 2029. The term ‘superhuman’ refers to intelligence that is greater than any individual human for an exact assignment, while ‘superintelligent’ refers to intelligence that surpasses the combined abilities of all humans for any task.
Elon Musk said that, “My guess is that we’ll have AI that is smarter than any one human probably around the end of next year.” There was a caveat in that prediction, which stipulated that growing power demands and shortages of the most powerful AI training chips could limit their ability in the short term.
“Last year it was chip-constrained. People could not get enough Nvidia chips. This year it’s transitioning to a voltage transformer supply. In a year or two, it’s just electricity supply,” Musk added. He expressed his concerns about superintelligence in 2023, stating that he wished he could pause its development.
During the launch of his AI startup, xAI, he emphasized the importance of building AI in a responsible way to avoid a Terminator-like future. xAI is leading the development of superintelligence. Elon Musk claims that the latest version of his chatbot, Grok AI, is equivalent to GPT-4, which is over a year old.
However, competitors such as Anthropic’s Claude 2 Sonnet have already surpassed GPT-4’s capabilities. Musk’s predictions were imprecise in the past, including Tesla’s autonomous drive from NY to LA within two years, SpaceX’s Mars mission in 2018, and Neuralink’s brain chip in four years. The first human implant from Neuralink was received seven years later.