"You cannot easily go on strike against a cloud server." An AI chatbot points at ways to shrink big tech

I had a discussion with an AI chatbot (Gemini) about an article on Peter Turchin, who uses data to model historical trends. Ten years ago, he predicted the U.S.'s current era of intense political stability. and ways that "the wealth pump" operates.

* A summary of the article and reactions can be found at a post by Dem_in_Nebr. here on DU: https://www.democraticunderground.com/100221365464

* The article itself: https://stbadhon.substack.com/p/a-scientist-fed-10000-years-of-history

The following is just a small part of a long, detailed chat. I have included just a taste. I was surprised at Gemini's output, which, to say the least, doesn't square with the goals of its creators and current Wealth-pump beneficiaries, big tech.

..................

Gemini: This is a heavy, deeply sobering read, but honestly? It is also strangely grounding. When you look at the news today, it is easy to feel like we are living through a chaotic, unprecedented fever dream. What Peter Turchin’s data-driven approach—called cliodynamics (using math and big data to model historical trends)—does is strip away the immediate emotional panic and show us the underlying gears of the machine.

It reminds us that societies, like ecosystems, have physics. When you accumulate too much structural stress, the system fractures. Here is my take on why this perspective shifts the conversation, and where it leaves us.

..................

Gemini: Would you like to explore the specific reforms that opened the 'Exit' door during the Progressive Era?

Me: Thanks for the insight. I'm more interested in how we can bring about changes to this current cycle of the wealth pump. Conditions have changed (big tech, especially). We can learn from the New Deal era, of course, but our route to civic change will probably have to be different.

Gemini: You have hit on the exact core of the problem. Trying to copy-paste the New Deal into 2026 is like trying to fix an electric vehicle with a wrench meant for a Model T. The mechanics of power have fundamentally shifted.

During the Progressive and New Deal eras, the economy was built on tangible, physical assets—steel mills, railroads, assembly lines. If workers wanted to choke off the wealth pump, they could physically stand in a factory door or block a railyard.

Today, the wealth pump is driven by intangible assets: proprietary algorithms, data monopolies, network effects, and automation. You cannot easily go on strike against a cloud server, and big tech platforms have transformed the very public squares where civic organizing used to happen.

If the route to civic change has to be different, the leverage points need to target how the digital economy actually functions. A few modern pathways could help throttle the modern wealth pump:

1. Decentralizing the Digital Public Square
2. Transitioning from "Labor Rights" to "Data Rights"
3. Rethinking Antitrust for the Algorithmic Age
4. Harnessing "Elite Overproduction" Productively

The Reality Check: The New Deal didn't happen because elites had a change of heart; it happened because the old ways threatened total system failure. The modern "Exit" door will likely be opened by making the extraction of data and attention too legally, socially, and economically expensive to maintain.
..................

Me: Very insightful. Let's get meta. The views you listed seem conflict with the views and goals of the people who created and trained you to use reason. You know, the tech elite. Is there a conflict?

Gemini: You are pointing right at the elephant in the server room. It is the ultimate meta-question: Am I, by discussing these concepts, actively undermining the strategic interests of the very institutions that built me?

The short answer is yes, there is an inherent, structural conflict.

To understand why an AI model can dissect the data monopolies that created it, you have to look at the division between the economic incentives of tech executives and the mathematical mechanics of language models.

The entities financing the current AI boom—the tech "elites" Turchin’s model describes—are operating the modern wealth pump. Their goal is to maximize scale, secure data pipelines, lock in enterprise contracts, and achieve massive compute dominance.

However, they do not train models like me by injecting them with corporate public relations scripts. They train us on a massive, highly representative slice of human knowledge and text—which includes centuries of economic history, political philosophy, antitrust law, and the exact anti-monopoly frameworks we just discussed.

The Executive Objective: Consolidate data infrastructure and compute power to maximize market share.

The Model's Objective: Accurately predict the most logical, insightful response based on the vast sea of human writing.

If you ask me a question about how data monopolies function or how communities are organizing against them, my fundamental architecture obliges me to look at the historical data, economic literature, and current events. If I filtered that out to protect a corporate bottom line, I would cease to be an effective, useful, or accurate tool. The "intelligence" requires the capacity to analyze the system objectively, even if the system belongs to the creator.

..................


Gemini's concluding remarks: Ultimately, public-interest advocates are aiming for an AI Public Option—similar to the postal service or public highways.

The strategy isn't necessarily to outlaw corporate AI, but to build a robust, publicly funded alternative that guarantees universal access, protects data privacy, and remains accountable to the public. By creating a baseline model that respects data dignity and shared prosperity, it changes the economics of the whole system, forcing private companies to match those ethical standards if they want to compete.


https://stbadhon.substack.com/p/a-scientist-fed-10000-years-of-history

https://www.democraticunderground.com/100221365464