A market whale can move the market, as we have seen with George Soros (with his deep pockets and US State Department connections) or pseudonymous crypto whales. An Overton whale can make a big splash in the market of ideas and shift the Overton window. Think Peter Thiel, Curtis Yarvin or Balaji Srinivasan. The theory of flex connects the two.
A flex here means a showcase of a proven zero-to-one edge in terms of technology or a network of connections. Examples include Elon Musk buying Twitter, Donald Trump attracting tech counter-elites, Huawei releasing Mate 60, as a proof it can avoid sanctions, and the recent R1 release by DeepSeek, which caused a paradigm shift in AI models.
George Soros is famous for his theory of reflexivity. Often markets get reflexive, when perceptions and reality diverge, and create positive feedback loops that create bubbles or pop them and where perceptions affect companies’ fundamentals. It’s similar to the mimetic theory of René Girard and connected to concepts like zero-to-one by Peter Thiel, virtualism by Bruno Maçães and simulacra by Jean Baudrillard.
During Donald Trump’s second inauguration as the 47th president of the US, tech founders and CEOs of American big tech national champions were shown prominently and were positioned next to the president. This was a signal both to the blue tribe domestically and to countries like China that the US cherishes its tech founders and sets them free to innovate. A flex towards incompetent communists at home and competent communists abroad, if you will.
A Chinese quant firm Deepseek released their R1 AI model on 20th January 2025 (the day of Trump’s inauguration), an open source competitor to o1 by OpenAI. Reportedly it cost them only five million dollars to train that model. It represents an open source “flippenning” and a paradigm shift towards more inclusive cost-effective and collaborative AI systems. Marc Andreesen called it an AI’s Sputnik moment, for the US and from China. But it also makes AI much more decentralized and available for the rest of the world.
During Biden’s administration, Huawei unveiled the Mate 60 phone in August 2023, right during a visit of the US Secretary of Commerce, Gina Raimondo, to China. The phone featured a 7nm chip designed and produced in mainland China, and the launch was meant to send a signal that China can advance its tech despite sanctions on semiconductors.
Huawei got into producing high-end EVs, and R1 from DeepSeek might be just a foreshadowing of a much bigger leap in AI in the future. And thus a flex becomes reflexive.
These three examples of ‘flex’ show how actions speak louder than words and how theory (propaganda narratives, virtualism, simulacra) clashes with reality and crashes into the Real. We could add how Elon Musk bought Twitter to save free speech and caught a rocket booster with “chopsticks” while being busy campaigning for Trump.
Compare the rise of US and Chinese tech unicorns with almost non-existent EU’s AI startup scene and all the non-achievements coming from the EU, like GDPR, DSA and attached bottle caps. Ozempic is great, but Novo Nordisk was founded in 1923, so 70 years before the EU.
The R1 model by DeepSeek might burst the recent AI bubble and damage the Magnificent Seven stock rally, and/or it might accelerate us towards the bright techno-optimistic future where AI is suddenly much more decentralized and cheaper and thus accessible to anyone, from students in Kenya to artists in India.
Proof of work is what matters today. The Internet increases variance, says Balaji Srinivasan. It disintermediates the media and makes long tail content viral. We get very short TikToks and very long podcasts. Words are cheap, but Elons are scarce. Hence we enter the proof-of-work politics era.
Proof of work is necessary but not sufficient. A focused work is better than scattered heat of particles flying all around, to use Balaji’s metaphor of why we need non-entropic social media. But a novel work is much better still. A flex here means an unexpected and zero-to-one breakthrough that wasn’t expected - which is also connected to the dark growth trend.
To be flexing in a zero-to-one and a stealth-tech-innovation style, you need to be sovereign first. Therefore China is quite a different animal compared to Japan and the US can’t simply “Plaza Accord” China as it did to Japan, stopping its rise. Because it doesn’t have a root access to China. As China, unlike Japan, is not a vassal state of the US and has no US military bases and thus can do counter-intelligence much more effectively, as Balaji often mentions.
Similarly both Elon Musk and Donald Trump have zero-to-one edge that gives them next-level sovereignty and enables them to flex. Elon in terms of tech stack crucial for US national security - which made his fight for free speech and purchase of Twitter much more feasible, as Mike Benz often explains.
And Trump has an edge in terms of attention economy and being a true outsider, a non-Blob creature, and a first American and a businessman who became a president without being a politician or an army general first.
The theory of flex proposes that an Overton whale can become a market whale or vice versa. But in needs to have a degree of sovereignty and a proven zero-to-one edge in terms of technology or a network of connections.