Strange bedfellows make for lazy headlines. When a democratic socialist from Vermont and an architect of nationalist populism lock arms to condemn the artificial intelligence barons, the media eats it up. The lazy consensus writes itself: big tech is creating a neo-feudal dystopia, a digital aristocracy hoarding automated wealth while the working class fights for scraps. It is a neat, emotionally satisfying narrative.
It is also completely wrong.
I have spent the last decade watching legacy systems panic over technological displacement, and the panic today is louder than ever. I have seen corporations blow millions on compliance theatre while missing the structural transformation right in front of them. The populist critique of artificial intelligence relies on a fundamental misunderstanding of how capital, compute, and open markets actually operate. Bernie Sanders and Steve Bannon want you to believe that artificial intelligence is a monolith of centralized control, a private club of modern robber barons building digital monopolies that can never be broken.
They are looking at the wrong century.
The Myth of the Unbreakable Moat
The core premise of the anti-oligarch argument is that machine learning models require such staggering amounts of capital and data that only a handful of trillion-dollar balance sheets can survive. Look at the server farms, they cry. Look at the multi-billion-dollar training runs. Look at the energy grid demands.
This argument treats computing power like nineteenth-century steel mills. It assumes that scale is a permanent barrier to entry. History screams the opposite.
Every single foundational wave of modern computing has followed the exact same trajectory: initial extreme centralization followed by hyper-commoditization. In the nineteen-sixties, mainframe computers were controlled by a tiny cartel of corporate giants. Mainstream commentators warned that humanity would be enslaved by centralized IBM behemoths. Then the personal computer arrived, followed by open-source software, followed by cloud infrastructure. The monopoly dissolved because the marginal cost of compute plummeted toward zero.
We are watching the exact same compression cycle happen at light speed. What cost fifty million dollars to train eighteen months ago can now be replicated on open-source weights for a fraction of that budget, running locally on hardware you can buy at a retail store. The proprietary moats that the political class loves to hyperventilate about are eroding faster than corporate legal teams can draft nondisclosure agreements.
The oligarchy does not fear regulation. They love it. Heavy-handed state intervention, pushed by populist politicians terrified of disruption, is the single greatest gift you could hand to incumbent technology giants. Compliance costs kill startups; they are merely a rounding error for conglomerates with armies of lawyers. When Sanders and Bannon demand state control or punitive antitrust dismantling under the banner of protecting the little guy, they are pulling the ladder up behind the incumbents.
Open Source is Eating the Cartel
Let us talk about where the real power is shifting. While Washington politicians posture about algorithmic overlords, the global developer community is quietly turning proprietary architecture into a public utility.
Open-weight models are the great equalizer. When Meta, Mistral, and thousands of independent researchers release models that match or beat closed systems, the entire premise of the corporate monopoly evaporates. You do not need permission from a Silicon Valley boardroom to deploy intelligence if you can run an elite model on your own local server.
The populist panic assumes that users are passive consumers, cattle being milked for data subscription fees. This ignores the decentralized nature of modern implementation. The real economic value is not locked inside the central training cluster; it is deployed at the edge, tailored by small businesses, local clinics, and independent creators solving hyper-specific regional problems.
I have watched local manufacturing outfits completely outmaneuver multinational competitors not by building foundational models, but by fine-tuning open-source weights on proprietary supply chain data. They did not wait for Bernie Sanders to nationalize anything. They did not ask Steve Bannon for permission to reclaim national sovereignty. They just downloaded the weights, plugged them into their workflows, and ate their competitors' lunch.
The Energy Scarecrow
Another favorite talking point of the anti-oligarch coalition is the energy apocalypse. We are told that server farms will boil the oceans, drain municipal grids, and leave ordinary citizens shivering in the dark so algorithms can generate photorealistic images of cats.
It is a great soundbite. It is also analytically bankrupt.
Energy consumption in computing has always followed Jevons Paradox: as efficiency increases, total consumption rises because the utility of the tool explodes. But treating data centers as parasitic energy vampires ignores the economic incentives driving the industry. The firms building these clusters are not spending billions on electricity bills out of sadism. They are investing heavily in next-generation nuclear, geothermal, and grid-stabilization technologies because cheap power is their ultimate bottleneck.
The private sector is financing energy innovation at a scale that public infrastructure projects have not matched in decades. If you want a green grid, you should be cheering for high-compute demand, because it forces the rapid deployment of clean energy generation that would otherwise sit trapped in regulatory purgatory for twenty years.
Stop Fighting the Last War
The populist outcry against artificial intelligence oligarchs is a symptom of a deeper intellectual laziness. It is easier to rail against shadowy elites than to understand the gritty mechanics of distributed systems, vector databases, and inference optimization.
When you frame technological progress as a class war between the sovereign state and the evil tech billionaire, you miss the actual battleground. The battleground is not about who owns the biggest server farm. It is about who has the operational agility to integrate intelligence into real-world workflows before their competitors do.
The next time you hear a politician lamenting the rise of digital feudalism, ask yourself what they are trying to protect. Usually, it is not the worker. It is the relevance of their own bureaucratic machinery in a world that is moving too fast for them to comprehend.
Do not wait for the politicians to save you from the oligarchs. Build something.
