Elon Musk's xAI aims to find out "what the hell is really going on"

Content

Summary

Updated July 15, 2023:

Elon Musk detailed his new AI project, xAI, in an audio event on Twitter. The goal, he said, is to create a super-intelligent AI that can solve complex scientific and mathematical problems.

During the Twitter event, Musk and his team of experts said their AI system should help society understand "what the hell is really going on." Musk believes it is realistic to expect such AGI to exist by 2029.

XAI, he said, is focused on deep scientific problems. Musk cited dark matter, gravity, and the Fermi paradox as areas of application. Musk noted that xAI is still in its infancy, but that it will work with Tesla and Twitter, and promised to protect investors' interests.

Tesla is also "by a wide margin" the leader in applying AI to the real world, Musk claimed, pointing to the self-driving features in Tesla cars.

Despite its strong scientific focus, Musk said xAI is also intended to compete with OpenAI and create offerings for consumers and businesses. News of xAI's first release is expected in a few weeks.

A recording of the show is available here.

Update from 12. July 2023:

Elon Musk's OpenAI competitor "xAI" now has a website

Elon Musk's new AI company, xAI, has officially launched with a mission to "understand the true nature of the universe," according to its new website. The xAI team, which includes experts who previously worked at OpenAI, Google Research, Microsoft Research, and DeepMind, is advised by Dan Hendrycks of the Center for AI Safety.

Recommendation

AI in practice

OpenAI's new "Orion" model reportedly shows small gains over GPT-4

Members have contributed to major breakthroughs such as AlphaStar, AlphaCode, Inception, Minerva, GPT-3.5 and GPT-4, Transformer-XL, Autoformalization, the Memorizing Transformer, Batch Size Scaling, and μTransfer.

xAI will be separate from Musk's X Corp, but will work closely with Tesla and other related companies. The company will host a Twitter Spaces discussion on July 14 for the public to meet the team and ask questions. You can find more information about xAI on its new website.

Update from 18. April 2023:

Elon Musk founds AI company X.AI and is working on "TruthGPT"

Elon Musk confirms work on "TruthGPT," a ChatGPT alternative that Musk says seeks "maximum truth" and "the nature of the universe." He made the announcement on Fox News.

Such an AI, Musk said, is the safest way forward because an AI interested in the universe would be unlikely to eradicate humanity as an interesting part of the universe - just as humans respect chimpanzees and would not exterminate them.

Musk also positions TruthGPT as a counter-movement to OpenAI, which is now "closed source" and heavily dependent on Microsoft, with Google and Deepmind as comparable commercial opponents, according to Musk. He said TruthGPT should become a third option, hopefully doing "more good than bad."

Article from 15. April 2023:

Elon Musk launches AI company X.AI in pursuit of OpenAI

OpenAI co-founder Elon Musk is getting back into the AI business. His new company is called X.AI.

X.AI is registered in Nevada, with Elon Musk as sole director. Jared Birchall, who runs Musk's family office, is registered as a secretary. As a private company, X.AI is authorized to sell 100 million shares.

"X" is Musk's vision of an "everything app," a single piece of software that can be used for all purposes of daily life and work, from social media to payments to, apparently, artificial intelligence. Musk also recently changed Twitter's corporate name to X Corp. and registered the company in Nevada. Its parent company is X Holdings Corp.

Rumors that Elon Musk is working on a ChatGPT alternative surfaced back in February. Musk has reportedly set up a research lab, with former OpenAI and Deepmind researcher Igor Babushkin in a leading role. According to Business Insider, Musk has purchased thousands of graphics cards, which are typically used to train large AI models.

OpenAI and Elon Musk share history

Elon Musk was one of the co-founders of OpenAI in 2015, and for a long time was its biggest backer. He stepped down from the board in 2018, officially to avoid conflicts of interest with his AI research at Tesla.

However, according to the US website Semafor, Musk reportedly failed in his attempt to completely take over OpenAI. Musk was unhappy with the development of the start-up, which he felt had fallen behind Google in AI development.

OpenAI's current CEO, Sam Altman, and other co-founders rejected Musk's offer, after which Musk left the company and stopped funding it. Musk reportedly invested $100 million of a planned $1 billion in OpenAI.

Elon Musk's love-hate relationship with AI

Since leaving OpenAI, Musk has been critical of the startup's development. He signed the AI Pause open letter, calling OpenAI's developments "extremely concerning" in terms of safety.

Musk is particularly concerned that OpenAI's AI chatbots are "too woke," or left-leaning. In mid-February, Musk called for "TruthGPT," a model whose name implies that it tells "the truth" from Musk's perspective, rather than, say, meeting certain public expectations.

Musk also criticized Microsoft's heavy involvement in OpenAI, which would indirectly be taken over by the software giant. OpenAI CEO Sam Altman responded to the Wall Street Journal's X.AI reveal on Twitter with an ironic "concerning".

Already in the years leading up to OpenAI's big breakthrough, Elon Musk has repeatedly expressed concern about the evolution of AI. In 2020, for example, he said that AI was on its way to becoming "vastly smarter than humans" and could reach that goal as early as 2025. From then on, he said, it would become "unstable or weird."

"My assessment about why A.I. is overlooked by very smart people is that very smart people do not think a computer can ever be as smart as they are," Musk said in an interview with The New York Times in the summer of 2020.

In 2017, Musk clashed with Facebook CEO Mark Zuckerberg when the latter called Musk an AI denier and doomsday prophet, prompting Musk to deny Zuckerberg's AI expertise. In the same year, Musk warned that developments in AI could lead to World War III.