In April Elon Musk told right-wing commentator Tucker Carlson that was the case start a project to compete with ChatGPT and “build a maximal truth-seeking AI trying to understand the nature of the universe.”
Today Musk presented this new project in the field of artificial intelligence. It’s called xAI. The company’s surrogate landing page reiterates this goal of understanding the universe, listing 11 AI researchers — seemingly all men — who have made significant contributions to the field of AI over the past several years, and at companies like Google, DeepMind, and OpenAI have worked.
According to Linxi “Jim” Fan, an AI researcher at Nvidia, the crew is an “all-star founding team.” “I’m really impressed by the concentration of talent – I’ve read too many articles by them to count,” he says writes in a LinkedIn post.
One of the company’s co-founders, Greg Wang, said in a tweet that xAI aims to take AI to the next level by developing a mathematical “theory of everything” for large neural networks, the machine learning technology that has dominated AI for the past decade. “This AI will enable anyone to understand our mathematical universe in ways previously unimaginable,” he wrote.
Like many other new AI projects, Musk’s concern and possibly some FOMO is motivated by ChatGPT’s rapid rise. He spoke of xAI in response to the bot, which he says has political bias, and criticized its creator, startup OpenAI, for being secretive and too comfortable with its supporter Microsoft.
Musk’s displeasure may be compounded by the fact that he co-founded OpenAI in 2015 but severed ties with the then-nonprofit three years later allegedly he fails to take full control. (The company became a for-profit in 2019.) And Musk recently joined in warnings that AI could represent an existential threat for mankind and consolidate power by giants like Microsoft and Google.
Musk is no stranger to making bold bets, but what little has been revealed about xAI’s goals sounds a little odd. ChatGPT and its competitors like Google’s Bard are based on deep learning, and Sam Altman, CEO of OpenAI, said that completely new ideas are needed to go beyond existing systems. Researching the basics of the technology might help find them.
However, much of the recent advances in AI have come as existing systems are getting larger and being given more processing power and data. And the sweeping changes that AI is expected to bring to technology and other industries over the next few years will come from the deployment of this largely mature technology.
As of this writing, xAI probably doesn’t seem to have the cloud computing power it needs to compete with OpenAI, Microsoft, and Google. And its relatively small team of AI researchers doesn’t look world-class compared to the hundreds each of these established companies can deploy to AI projects. The only person involved who has worked on AI risks in the past is xAI’s only named advisor, Dan Hendrycks, who is director of the nonprofit Center for AI Safety and recently published a public statement from technology leaders about the existential threat posed AI could represent coordinated.
Of course, while his supposedly giant-killing AI project starts small, Musk has some significant resources to draw on. According to the xAI website, the new company will work closely with Twitter and Tesla. Twitter’s data from conversations on the platform is well suited for training large language models like the one behind ChatGPT, and Tesla is now developing its own custom AI chips and has extensive experience building large compute clusters for AI to boost cloud computing -Performance of xAI could be used. Tesla is also building a humanoid robot, a project that could be supported by xAI in the future and could be helpful.
But perhaps at this early stage, xAI’s reality-warping rhetoric is all about attracting talent. AI expertise has never been so in demand. The most pressing problem for a new entrant, even one with Musk’s reputation and financial resources, is showing that it can attract the researchers it will take to eventually become competitive.
The big goals Musk has set himself – to challenge existing AI giants and protect humanity from malicious AI – make his little new AI company seem even smaller. Many AI researchers, who are also concerned about AI development, seem to view the problem as one that requires more transparency and collaboration, rather than a lone genius with a small group of all-stars.