Elon Musk has a knack for surprising the world with his bold and ambitious moves but his newest venture, artificial intelligence company xAI, may come as less of a surprise.
The launch sheds some new light on one of Musk's more puzzling and controversial decisions: his acquisition of Twitter.
Musk bought Twitter in October last year, naming the new holding company 'X Corp', with a vision of the social media company becoming an "everything app".
The billionaire tech baron said the new company aims to "understand the true nature of the universe", a lofty goal with echoes of fictional supercomputer Deep Thought from book series The Hitchhiker's Guide to the Galaxy.
The deal was met with scepticism and criticism from many quarters, as critics questioned Musk's motives and intentions, amid fears that he would use the platform to promote his own agenda and interests, or censor and manipulate its users and content.
It was unclear how Twitter would fit into Musk's existing portfolio of companies, such as Tesla, SpaceX, Neuralink, and The Boring Company.
The launch of xAI may provide some answers to these questions – though much scepticism still remains.
xAI is a separate entity from X Corp, but the AI company will work closely with the holding company, as well as with Tesla and other companies owned by Musk, and will leverage Twitter's data.
This data – arguably the largest and most diverse archive of human chat and communication ever created – is key to Musk's latest commercial offering.
In the past week Musk has sought to protect the exclusivity of Twitter's data through recent tweet-viewing limits that he said would block the 'scraping' of the entire database by potential rivals.
AI rivals OpenAI and Meta have been sued by content owners such as Sarah Silverman, and authors Christopher Golden and Richard Kadrey, over claims of copyright infringement, on the basis that the generative AI bots were allegedly trained on illegally acquired datasets.
Twitter's data therefore becomes a unique is a goldmine for generative AI, which – if you have been hiding in a cave – is a branch of AI that can create new content such as text, images, audio, and video by being 'trained' using huge volumes of data, including articles, books, images and tweets to generate new content that mimics the style and content of the original data.
Musk is no stranger to generative AI. He became a co-founder of OpenAI in 2015, but left the nonprofit in 2018 over disagreements with its direction and vision.
He has since been critical of OpenAI's products, such as ChatGPT and image-creating model DALL-E.
In April he co-founded xAI, hinting that he was working on developing an alternative to ChatGPT called “TruthGPT”, which he said in an interview would be a "maximum-truth-seeking AI" that would use Twitter's data to generate factual and unbiased content.
With xAI, Musk is building on years of tech ambitions and utilising his contacts in the industry. The company has a team of veterans from DeepMind, Google Research, Microsoft Research and Tesla, which have together contributed to some of the most widely used methods and breakthroughs in the field, such as AlphaStar, AlphaCode, Inception, Minerva, GPT-3.5, and GPT-4.
The future of the tech could be anything from exploring fundamental physics to creating digital simulations. Musk has not yet revealed much else about the startup's plans and projects.
A Twitter Spaces session will be hosted by Musk on the platform on Friday where users will be able to ask questions.
The launch of xAI may not fully justify or explain Musk's acquisition of Twitter, but it does show that he has a vision and a strategy for using and commercialising its data and platform.
It also shows that his competitive juices are flowing as strongly as ever, with the South Africa-born entrepreneur recently stung by Mark Zuckerburg's launch of the 'Twitter killer' Threads social media platform.
Indeed, Musk is far from alone in leveraging existing social media resources in the AI race.
Zuckerburg's Meta Platforms is also, it has been revealed today, planning to launch a 'ChatGPT killer', which will be able to leverage a global dataset of many billions of user posts, comments, and reactions on Facebook, Instagram, Threads and perhaps even WhatsApp.
Each tech titan exploring the realm of AI has a unique vault of data at their disposal and their success may be dictated by their ability to extract and utilise insights from these individual datasets.
Google's Bard project can tap into the expansive history of the internet that the Alphabet Inc (NASDAQ:GOOG) company has indexed over many years.
Simultaneously, Microsoft's partnership with OpenAI has access to a wealth of data from a mix of professional, social, and gaming interactions across Microsoft's platforms, providing a diverse pool of information that could result in an AI with a robust understanding of multiple domains.
Whatever their various and ultimate aims, the ensuing developments will significantly shape the future of AI and possibly the planet – and the race is seemingly only just beginning.