Much ink and many pixels have been spilled on Twitter’s gradual rebranding to an X logo. Most of it argues that Elon Musk will fail since the social network numbers and reliability are down, hate speech has gone up (Twitter denies this), and the service now faces its first real competition from Facebook/Meta. I would argue that these are all the wrong metrics.
The real target is taking over the front door to Internet services now occupied by Apple, Google, Uber, and Airbnb, hence the term X or everything. Previously diginomica has published a guide to Everything-as-a-Service (XaaS) for businesses. Musk is attempting to extend this concept to a consumer platform.
Musk does have a checkered history of projects and businesses that failed to deliver as promised. The Boring company hasn’t reduced traffic, the hyperloop has been passed off to others, and regulators are starting to complain about the marketing of “full-self driving,” which still requires hands on the wheel at all times.
But it’s important to appreciate that Musk has also built the world’s most efficient car factories while legacy car makers are struggling to transition to electric vehicles. He also built the largest ecosystem of electric car chargers in the US, built a space platform that launches ten times as many rockets at one-tenth the cost as American competitors and launched a fleet of over 3,900 Starlink communications satellites out of 6,700 total operational satellites.
So is X a boring social media company or a new platform that follows in the footsteps of Tesla, SpaceX, and Starlink? In the long term, it is more likely to become the latter if only because it is becoming a dominant focus of his attention, and also it follows in the footsteps of his earliest vision for PayPal before the cloud infrastructure, compute, and AI ecosystems were in place to support it.
Catching up to Google lead
Today, Google has a clear lead on the search engine gateway to the Internet with over 92% of the market compared to 2.76% for Microsoft’s Bing, and that’s after the launch of the Bing Chat large language model (LLM) service. Google is also considerably larger than most other Internet platforms, with over $284 billion in revenue, compared to $116 billion for Facebook, $32 for Uber and $4 billion for Twitter in 2022. Google also has a lead in location services.
But it has also stumbled in other important areas. For one, Google has struggled with a dizzying array of voice, text and video communications apps that sometimes communicate with each other and have often been cancelled. But more significantly, it never took advantage of its early lead in pioneering the Transformer AI architecture in 2017, which ushered in LLMs and now dominates much of the recent hype in AI. It was not until the recent success of ChatGPT that it prioritized the new field. And that was after all of the researchers who wrote the pioneering paper Attention Is All You Need went on to other companies or started their own ventures.
- Aidan Gomez co-founded Cohere for big data LLMs
- Noam Shazeer co-founded Character.ai for chatbots.
- Ashish Vaswani and Niki Parmar co-founded Essential AI for businesses,
- Lukasz Kaiser left for OpenAI.
- Jakob Uszkoreit co-founded Inceptive for biotech
- Illia Polosukhin co-founded NEAR to create a blockchain operating system.
- Llion Jones recently left to launch a secretive startup in Japan.
In the meantime, Elon Musk has recently launched X.AI to develop a coherent AI platform, which will undoubtedly underpin the transition from a social media company to an everything app.
In the long term, digital twins will play a growing role in all this. This month, Tesla plans to turn on its new Dojo Supercomputer, purpose-built for improving its autopilot. The first goal will be building a more reliable self-driving service. Eventually, it could also provide a constantly updated digital twin fully connected to the X platform.
It’s also significant that its full-on map is being created by Tesla drivers that have paid for the cars and, in some cases, up to $15,000 for the full self-driving package Companies like Nexar have been working on a service to build constantly updated digital twins since 2021 on top of dashcam footage. Google reportedly amassed 10 million miles of curated Street View imagery by 2019. As of May 2023, the Tesla fleet had exceeded 100 billion miles.
But Tesla and now X never had the infrastructure to process this data at scale. The new effort to build a better everything app, supporting AI infrastructure, and Dojo could change that.
Where are the super apps?
Super apps, like Alipay, WeChat in China, PayTM and Tata Neu in India, and GoTo in Indonesia, are all the rage in Asia, but have failed to catch on in the US. They rose to prominence since they filled a void left vacant by the existing payment and banking infrastructure.
In the West, similar apps have had to compete with a working and efficient banking infrastructure, not to mention a 30% tax imposed by mobile platforms like Apple and Google. Twitter CEO Linda Yaccarino believes they can change this situation. On Sunday, Yaccarino tweeted (or is it now X-ed?):
X is the future state of unlimited interactivity – centred in audio, video, messaging, payments/banking – creating a global marketplace for ideas, goods, services, and opportunities
It’s tempting to question Musk’s insistence that he is building TruthGPT underpinning this emerging X-ecosystem as a “maximum truth-seeking AI.” He has a habit of spreading half-baked ideas on his favorite social medial platform that don’t hold up to the light of day or that result in lawsuits or Government oversight.
And despite the bluster, I earnestly believe he wants to pursue the truth wherever it may lead. After downplaying the need to halt production during COVID, he launched a comprehensive PCR testing program at SpaceX. This resulted in some groundbreaking research on the role of antibodies in protecting against the disease published in Nature.
Tesla also has a leg up on building a world scale digital twin that plugs into other services. This could better maps, constant updates on road conditions, new enterprise services, and a massive data set for Apple’s VisionOS.
He also has demonstrated a series of unlikely successes in approaching established markets in autos, space, satellites, and payments with a platform mindset that incumbents offered. It is sure to be an uphill battle. But as venture capitalist Marc Andreessen recently said:
I have developed a view in life not to second guess Elon Musk.