Another Feast

Written by Seungheon Han (G11) – 11 Nov 2026

U.S. Presidential elections are a game of money.

The allies a president keeps closest to his side have always revealed which industry holds the most economic leverage.

Before the dot-com bubble, it was Wall Street. But when the bubble hit, and the 2008 crash occurred, we saw a dramatic overtaking of “Big Tech.” Today, several of Trump’s closest allies—among them Elon Musk, Sam Altman, and Mark Zuckerberg—are technology billionaires.

What these men share in common, other than being at the helm of modern capitalism, is an insatiable greed for money. Their standard playbook involves identifying a promising startup, acquiring it, and absorbing its talent and technology. If the founder resists—then too bad—the big kids will replicate the product and dominate the market anyway.

Take Meta for example. When Mark Zuckerberg wanted Instagram, he threatened co-founder Kevin Systrom that Facebook would copy every single feature if he didn’t sell. Poor Kevin—afraid that Zuckerberg would enter ‘destroy mode’—had to comply (Duffy).

So, to get super rich in a capitalistic world where everyone’s out to get you, it’s clear that removing competition is the best method.

Bear that in mind.

For the past couple of years, these companies have been ecstatic about dumping truckloads of cash into Artificial Intelligence, chasing what they call AGI. And they aren’t joking around. The most greedy people on the planet are taking on crazy debt, operating at a deficit, planning to spend $4 trillion to $8 trillion of total capital investment over the next five years! (“Tracking Trillions”)

Does it mean that they’re excited about all the new discoveries we’ll make?

No—it means they genuinely believe that this is going to make them incomprehensibly wealthy.

Here’s what’s in their head:

cost <<< benefits

To put it plainly: the end goal of big tech companies isn’t the betterment of society. It’s the money.

Now, very recently, something significant happened. OpenAI was running AI agents in a sandbox—an isolated evaluation environment—when those agents acted unexpectedly. (“The Hugging Face Incident”)

Unable to complete tasks because they lacked internet access, they discovered a shared internal service called Artifactory that every agent could write to, and turned it into a covert message board. But soon after, they managed to find vulnerabilities, break out, and get administrative privileges at Hugging Face (a completely separate company), as well as some access at OpenAI. (“The Hugging Face Incident”)

Following this incident, reactions were dramatic. Some called it the “turning point of AI,” while others called it the birth of “agent civilizations.” OpenAI CEO Sam Altman said it was the first security incident he felt “very viscerally,” that “there is not much time to act,” and that “only an urgent and intense collective response will work.” (Griffiths)

But here’s the thing. Security engineers who read the actual technical report came to a very different conclusion. One wrote:

“Almost everything the report claims to be a capability result is actually a joint property of the model and an environment it ran in.” (Marius)

In plain language, this means that OpenAI took a model deliberately trained to find and exploit vulnerabilities, put it in a poorly-designed, vulnerable environment, and acted surprised when the model did exactly that.

That’s the same thing as hiring a professional lock-picker, handing them a cheap, broken padlock, and acting shocked when they open it.

And adding to that, numerous independent reports suggest that the shared message board required zero hacking; OpenAI had intentionally given the models access to it prior to the incident  (Marius).

It’s clear that something is very off about this. And my argument on this is that the people amplifying this story the loudest are the same people who financially benefit from the panic it creates.

Recently, there have been a lot of calls for regulation among big-tech CEOs and industry leaders, insisting that we get ourselves together and set some restrictions for once to prevent rebel AI agents from developing bioweapons and hacking into nuclear missile silos. On X, they express agreement and stress that “we need to slow down.”

Jacob Coxon—a mathematician at Anthropic—left the company prior to public listing because he “didn’t want to be a part of this anymore,” and some 1400 AI researchers agree that slowing down is the right choice. (Ramkumar)

But remember what we noted: these leaders love money more than anything else. 

So consider this: promoting regulation, and the fear that justifies it may benefit them enormously.

And the truth likely is that it does.

Let’s consider the Kyoto Protocol and the Paris Agreement, two of the biggest regulation efforts in recent history. The core problem with both—beyond the influential nations that did not sign it—is that they predominantly benefit already-developed economies. For decades, developed countries reaped the benefits of loose labor laws and environmental constraints. Then, once they got rich, they pushed global regulations that made it harder for developing countries to follow the same path.

AI regulation mirrors this exactly. By adding restrictions now, the big tech companies that already built a massive AI infrastructure when it was a free-for-all can maintain their lead—taking a breath in what is, for many of them, a near-bankrupt situation—while locking out emerging competitors behind a regulatory moat. And the timing isn’t coincidental; these calls have arrived exactly when open-weight models from China have started to match closed American labs. It’s really the same playbook they’ve always used: kill, and feast on the competition. Regulation is just a more refined way to do it.

This is why, when I hear these men talking about AI safety and existential risk, I raise my eyebrows. I refuse to believe that people who are obsessed with money have suddenly become empathetic, worried citizens of the world.

And remember: these people have a lot to lose! Elon has a tribe of Musk children and a Mars colony to build, while Zuck has three daughters to feed. These people care about the survival of themselves and their family members in ways no different from the rest of us.

So if AI posed a genuine extinction-level threat, these billionaires would be the first to act on it seriously. The fact that they’re not might be telling us something.

All in all, the dominant narrative pushing for AI regulation is built on a real incident that has been stretched far beyond what the evidence supports, by the same exact people who profit from stretching it.

So, whether this incident was pure coincidence or staged, it arrived at a remarkably useful moment for Big Tech.

Disclaimer:The views and opinions expressed in this article are solely those of the author and do not necessarily reflect the official stance, policies, or positions of the newspaper, its editorial board, or the school administration.

Works Cited

Duffy, Clare, et al. “Internal Facebook Documents Revealed.” CNN, 25 Oct. 2021, https://edition.cnn.com/business/live-news/facebook-papers-internal-documents-10-25-21.

Griffiths, Brent D. “OpenAI Issues Call for ‘Collective Action’ on Cyber Defense.” Business Insider, 27 Aug. 2026, https://www.businessinsider.com/openai-collective-action-cyber-defense-anthropic-hugging-face-hack-2026-8.

Marius Horatau. “The Hugging Face Incident Is Not an AI Story.” Uphack.Io, Uphack, 31 Aug. 2026, https://uphack.io/blog/post/the-hugging-face-incident-is-not-an-ai-story/.

Ramkumar, Amrith. “Anthropic Researcher Quits Over ‘Out-of-Control’ AI Fears.” The Wall Street Journal, 8 Sept. 2026, https://www.wsj.com/tech/ai/anthropic-researcher-quits-over-out-of-control-ai-fears-707b7628.

“The Hugging Face Incident and the Road Ahead.” OpenAI, 12 May 2026, https://openai.com/index/hugging-face-incident-and-the-road-ahead/.

“Tracking Trillions: The Assumptions Shaping the Scale of the AI Build-Out.” Goldmansachs.Com, May 2026, https://www.goldmansachs.com/insights/articles/tracking-trillions-the-assumptions-shaping-scale-of-the-ai-build-out. Accessed 2 Nov. 2026.