Elon Musk Sets out to build OpenAI Rival

The company he helped found is now the enemy.

Welcome Back!

I wanted to report on Elon Musk’s incorporation on March 9th, 2023, a couple of weeks before the 6-month moratorium petition, the founding of X.AI. This is the future rival of OpenAI led by Elon Musk and that has already lined up investors from SpaceX, Tesla and other participants.

You can Restack my note here:

I wanted to go into some detail and give you my take on this event. Since it may be important for the future events related to A.I, AGI and the deep-seated technological centralization now occurring in American monopoly capitalism.

The Verge article was recently trending on Artifact. Musk is one of the co-founders of OpenAI, which was started as a non-profit in 2015. Fast forward eight years later and ChatGPT went viral, and this appears to have angered Musk.

It was an open secret that Musk was planning to build an OpenAI rival even as Google (via project Gemini), Anthropic AI, Aleph Alpha and AI21 appear hot on the heels of OpenAI and their foundational models. Meta also will build a decent rival, and there are many others.

If you enjoy any part of this article, you can highlight it and “Restack” is on Substack Notes.

Elon Musk the Sci-Fi Dreamer

Elon Musk has promised us everything from colonies on Mars to General Purpose Robots, SpaceX’s Starship launch is happening soon. The launch window is scheduled to start at 7 a.m. CDT (12 UTC) on April 17, 2023, and is open for approximately three hours. In the end, Twitter may go bankrupt but Elon Musk seems as powerful as ever both in the media and trusted by the Venture Capital community of Silicon Valley.

Musk advocated for the technology’s regulation and signed an open letter that urged tech leaders to put a six-month pause on its development, but appears to be leveraging both politics and A.I. anxiety to buy him more time to catch up.

Igor Babuschkin who has both experience at OpenAI and DeepMind may be involved in the project. Earlier this year SpaceX raised $750 million in a new round of funding that values the company at $137 billion, many believe SpaceX has a better long-term future than Tesla, which will face significant competition in the EV space soon.

With ChatGPT likely to face regulatory scrutiny both in the U.S. and in Europe, the direct competitors of OpenAI may have a window to scale up their foundational models, chatbots and spring into action with GPT Agents of various kinds. Elon Musk as the owner of Twitter could train his models on its proprietary data and come to some interesting results.

Rumors about Musk starting up an AI company have realistically been showing subtle signs not for days, but for weeks. Insider had reported that Twitter has bought “thousands of GPUs”. Elon Musk’s own criticism of OpenAI and seeming frustration was a tell tale sign of what was to come.

Must will likely try to use Generative A.I. to help him build his dream of an “Everything App” similar to China’s WeChat. It’s not clear how serious Tesla is with the Optimus robot or how all of this comes together, and if Twitter’s cash burning is really anywhere near under control, like he has seemed to claim in his recent BBC interview. The bizarre interview with the BBC had several million viewers and now has nearly one million YouTube Views:

Elon Musk, the second richest individual on the planet, with Tesla, SpaceX, Twitter, Neuralink, and other companies, and finally more than twenty years after founding SpaceX, X.AI Corp is born.

One of his children he named "X Æ A-12," pronounced "X Ash A Twelve”. The name was later changed to X Æ A-Xii. So why does Elon Musk seems obsessed with X? On July 5, 2017, Musk repurchased the domain name X.com from PayPal. It apparently has sentimental value for him.

Most believe X.AI Corp will become a direct rival to OpenAI. In fact, you could make the case that Elon Musk was ousted from OpenAI back in 2018 and has wanted back into the A.I. game every since. Rumors and circumstantial evidence of Musk building an AI company to pit against OpenAI have been around pretty much since he divested from the company — which he helped fund in its early days — back in 2018. Semafor’s coverage of why he split from OpenAI remains a classic reading for me in 2023.

But in early 2018, Musk told Sam Altman, another OpenAI founder, that he believed the venture had fallen fatally behind Google, people familiar with the matter said.

And Musk proposed a possible solution: He would take control of OpenAI and run it himself. - Semafor article

Five fateful years later, it’s ironically Google and its Bard that seem behind OpenAI, and Elon Musk had to sit on the sidelines becoming more critical of A.I. and more recently OpenAI itself. Suffice to say, Sam Altman and OpenAI’s other founders had rejected Musk’s proposal. With AI.X Corp, Musk can once again have his dream.

The latest development of mid April, 2023, reported by the Financial Times today, is that Musk has hired former DeepMind employee Igor Babuschkin and a handful of other engineers to begin work and has a number of his high-net worth individuals friends (read Billionaires) to help him fund the project.

Elon Musk is not just becoming more political, he believes he’s the one best able or capable, to make A.I. more safe and potentially prevent existential catastrophe. Curiously, a few days before Elon actually incorporated AI.X, he had criticized ChatGPT for its alleged political bias, describing the artificial intelligence chatbot as “too woke.” Back in December, 2022 Musk had previously criticized the technology, saying that "the danger of training AI to be woke – in other words, lie – is deadly."

Given the stockpile of A.I. talent OpenAI has and is now able to attract, and the level of sophistication of Microsoft Research, DeepMind, Meta AI and Google Brain, it’s a bit difficult to conceive X.AI will be able to catch up to GPT-5 type levels at the “speed of A.I.” or catch up at all very soon.

X.AI is incorporated in Nevada with Elon Musk as sole director. Elon Musk seems to want to build A.I. on his own terms. However competition with OpenAI seems to be heating up in 2023, even as Sam Altman is saying GPT-5 is not currently being trained and won’t be for some time. If the FTC is forcing OpenAI to halt GPT-5 training or if Microsoft has ascertained that it’s becoming too great a brand risk to rush things, there might be a renewed consolidation of the sector in 2024 and 2025 which will enable us all see who else comes out of stealth to show up to race and build.

A.I. workers in BigTech are not infinite. Here are where 33,000 of them currently work:

Tweets of Musk being openly oppositional towards OpenAI are not difficult to find. Clearly Sam Altman was the chosen one, and Elon perhaps became disgruntled when OpenAI decided to go down the corporate path to monetization and abandon their initial ideals.

Elon Musk is 51 years old, he’s approaching a time of his life where he may be concerned about establishing his legacy. While he’s driven Neuralink and Twitter staff hard in recent months, there’s not much evidence either of those projects will turn out well even as Tesla will face such significant competition that it might shock many retail investors and people who don’t understand the automobile business.

His views on A.I. have always been cryptic and a bit like an echo of the late Stephen Hawkings. “One of the biggest risks to the future of civilization is AI,” Elon Musk told attendees at the World Government Summit in Dubai, United Arab Emirates in February, 2023. So why is he now focusing on building a GPT-5 competitor that he has been so adamant about warning civilization about?

In a 2018 appearance on the “Joe Rogan Experience,” Musk teased his company Neuralink has something exciting in store for us. He believes his technology will allow humans to achieve a state of “symbiosis” with AI, where we'll be able to effortlessly combine our brains with computers. This sort of Silicon Valley transhumanistic thinking is not uncommon. But exactly this Silicon Valley bias also is directing how A.I. is developing potentially along unsafe grounds.

It’s not clear what kind of A.I. Elon wishes to merge with before he dies?

Elon Musk: Humans must merge with machines or become irrelevant in AI age

In 2017, he parroted much the same thing at the same conference:

In an age when AI threatens to become widespread, humans would be useless, so there’s a need to merge with machines, according to Musk.

Elon Musk has become more controversial in recent years, even as his wealth grew. How he seeks to harness the power of A.I. might itself become dangerous like a Silicon Valley tycoon’s own self-prophecy. Even for an Elon Musk fan, his bitter takeover of Twitter has been a bit unhinged (Even for Elon). He appears to have grown more ruthless with age.

I worry that Elon Musk’s financial power may corrupt his own common sense with how he plans to harness Transformers and Generative A.I. as a GPT (general purpose technology) for profit across the spectrum of his business portfolio and X Corp conglomerate. The centralization of power in technology usually does not have a good ending.

I would not wish to experience a ghost in the machine like Elon Musk.