---
title: "What’s really behind the AI panic."
type: "post"
post_id: "12665"
slug: "whats-really-behind-the-ai-panic"
canonical: "https://bombthrower.com/whats-really-behind-the-ai-panic/"
markdown_url: "https://bombthrower.com/whats-really-behind-the-ai-panic.md"
json_url: "https://bombthrower.com/whats-really-behind-the-ai-panic.json"
txt_url: "https://bombthrower.com/whats-really-behind-the-ai-panic.txt"
published: "2026-09-28T16:12:47+00:00"
modified: "2026-09-28T16:51:21+00:00"
author: "Mark E. Jeftovic"
categories:
  - "Zeitgeist"
tags:
site_name: "Mark E. Jeftovic is The Bombthrower"
publisher: ""
language: "en-US"
generator: "easyPress AI Visibility"
generator_version: "2.0.9"
---
![](https://bombthrower.com/wp-content/uploads/2026/09/ai-doom-brought-to-you-by.png)

The “Effective Altruists” want you scared, stupid and subservient.
------------------------------------------------------------------

*Today’s post is a sneak preview of the “Future Shock” section of the forthcoming Sovereign Capitalist Letter ([get our latest report and a sample issue here](https://thesovereigncapitalist.io/near-report-bonus/)).*

Just heading into mid-September, a purported moral panic broke out amongst the frontier AI leaders that their own products were “too dangerous” and so the only remedy would be for the government to regulate *all AI* models.

Weeks earlier, Bernie Sanders had introduced **The Ban AI-Superintelligence Act,** that would make it *illegal* to create an AI model that was more intelligent than humans, going so far as to prescribe a 20-year prison term for any *person* who did and the “corporate death penalty” for companies.

> JUST IN: 🇺🇸 Senator Bernie Sanders introduces bill to permanently ban AI that exceeds human intelligence.
> 
> Violators would face up to 20 years in prison. [pic.twitter.com/y2pk79zQlQ](https://t.co/y2pk79zQlQ)
> 
> — Watcher.Guru (@WatcherGuru) [September 3, 2026](https://x.com/WatcherGuru/status/2095567788051984638?ref_src=twsrc%5Etfw)

Sanders’ Bill is truly the domain of dystopian sci-fi. It reminded me of an episode from the ill-fated Twilight Zone reboot (“[Examination Day](https://en.wikipedia.org/wiki/Examination_Day)“) which culminated in a distraught couple mourning their child, who had just been exterminated by the state for testing too high on his pre-adolescent IQ test.

> Examination Day. [pic.twitter.com/PrKGnh7y4x](https://t.co/PrKGnh7y4x)
> 
> — Mark E. Jeftovic (@jeftovic) [September 27, 2026](https://x.com/jeftovic/status/2104324379479773592?ref_src=twsrc%5Etfw)

It all started with a[ “resignation in public” tweet](https://x.com/hilbertspaess/status/2097476196791709843) from one Jacob Coxon, an Anthropic employee, on September 8th posted his parting thoughts on X, stating that AI could *exterminate humanity within 10 years*.

> I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
> 
> — Jacob Coxon (@hilbertspaess) [September 9, 2026](https://x.com/hilbertspaess/status/2097476196791709843?ref_src=twsrc%5Etfw)

One of Anthropic’s trust and safety leads, Evan Hubinger, [then endorsed the tweet](https://x.com/EvanHub/status/2097497037956891126), citing a “non-zero chance” of 10% or more that such an outcome was possible.

That original resignation tweet, now closing in on *200 million views,* which emanated from a heretofore unknown X account, less than a year old, with no followers, and two name changes in the last six months went absolutely viral – with some well-timed, fortuitous boosting from some of the biggest “AI Alignment” personalities on the internet, and then the Wall Street Journal came out with an “exclusive” on the story, ….which actually went live on the WSJ site 18 minutes *before* Coxon’s tweet.

Before long [Sam Altman](https://x.com/sama/status/1635136281952026625) said we “definitely” need more government regulation on AI and even Elon Musk [chimed in and agreed ](https://x.com/elonmusk/status/2099026931006181867)that the frontier labs should be reviewing each other’s models for safety.

Within days a full-blown hysteria was underway – for me it bore all the hallmarks of *mass formation psychosis* or at least a well-organized attempt to manufacture one.

Within a week, King Charles was speaking at an AI Safety summit at Buckingham Palace, urging government oversight and increased regulation of AI…

![](https://bombthrower.com/wp-content/uploads/2026/09/king-charles-AI-Scotland-summit.png)
*“Surely we need sufficient means of control before it is all too late?”* the King ruminated.

Also, that guy behind him, looking rather pleased with himself, is “Tino” Cuéllar, Anthropic’s Chief Global Affairs Officer.

The “AI Safety Summit” itself was announced on September 7th, [leaked via Politico](https://www.politico.eu/article/king-charles-to-host-exclusive-ai-gathering/) – the day before the Coxon tweet that ignited what had become at this point, a global(ist?) furor.
Even the Pope [weighed on AI safety](https://x.com/cspan/status/2103508441759203437), and not to be left behind, Bill Gates [hit the Sunday talk circuit this past weekend](https://www.zerohedge.com/geopolitical/bill-gates-predicts-billion-deaths-evolutionary-event) with another doozy of a soundbite:

*“AI is powerful enough to cause a billion deaths” –* which means we need to turn control of it over to politicians and a new class of global inspectors.”

Several things about this relatively sudden AI panic did not pass the smell test, among them, the facts emerging around Coxon himself, who had “barely been at the company long enough to find the bathroom” (as the All-In guys put it), turned out to be neck-deep in the “**Effective Altruism**” movement, a well-connected network of young WEF types, which espoused global governance over potentially harmful technologies like AI.

He had received scholarships and grants from the EA Infrastructure Fund, the Long Term Future Fund and was a fellow at the Newspeak House in London (yes, *newspeak*).

**Who are the Effective Altruists?** 
-------------------------------------

Effective Altruism (EA) is a movement that purports to *“use evidence and quantitative reasoning to identify the most consequential problems and direct resources toward reducing them”.*

> Tired: World Economic Forum
> Wired: Effective Altruism [pic.twitter.com/zNGNfMNWkW](https://t.co/zNGNfMNWkW)
> 
> — Mark E. Jeftovic (@jeftovic) [September 16, 2026](https://x.com/jeftovic/status/2100230179259883521?ref_src=twsrc%5Etfw)

Love the euphemistic language (it’s actually what ChatGPT will use when you ask it). The punch line for the EA raison d’être is, as usual, a society ordered under expert, technocratic management – *their expertise, their management.*

One of their earlier recruits was Sam Bankman-Fried, whose FTX empire was run by a cadre of EAs at the top and who seed-funded Anthropic as the lead investor with $500M of a $580M Series B round in 2021.

A recurring EA strand is its fixation on existential risk from advanced AI, advocating stronger safety research, *governance*, and *oversight* to reduce the chance of catastrophic or irreversible outcomes.

We’ve all seen this movie before, with global warming, then with Bitcoin mining (causing more global warming) and now, AI exterminating humanity.

The common through-line is that the “problem” is some unfalsifiable calamity that *could* happen someday and for which the only logical response is to put a bunch of globalist technocrats in charge of… well, *everything.*

What people didn’t seem to understand, then or now – is that the problem isn’t really about the problem – I’ve long reported on proposals that [would make personal choice and individual liberty things of the past](https://bombthrower.com/bbc-climate-change-too-important-to-be-left-to-personal-choice/), because letting people think for themselves would be “too dangerous” for the problem at hand.

Global warming [meant personal carbon quotas](https://bombthrower.com/hear-me-out-personal-carbon-allowances-as-sound-money/), [limits on travel](https://bombthrower.com/coming-soon-your-travel-will-be-restricted-by-personal-carbon-allowances/), even living space, according to some “degrowth” advocates, how many new outfits you could buy.

> listening to greenoid Ulrike Herrmann sell her degrowth plan for Germany: no personal automobiles ofc, living space to be rationed at 50 sq metres/ person, train speed capped at 100 km/hour & train travel also to be rationed (everyone will get a personal kilometre allotment). [pic.twitter.com/JTlTlVDwVm](https://t.co/JTlTlVDwVm)
> 
> — eugyppius (@eugyppius1) [October 30, 2024](https://x.com/eugyppius1/status/1851654144089633138?ref_src=twsrc%5Etfw)

The panic around Bitcoin mining wasn’t about that, it was about *energy* – and [who gets to decide how much *you’re* allowed to consume](https://bombthrower.com/the-bitcoin-energy-debate-is-one-of-freedom-vs-servitude/) and what *you* can use it for.

Bitcoin mining today, Netflix tomorrow. Why not? Once you concede oversight to some external authority for one form of energy use, you’ve tacitly done so for all of them.

The problem that politicians and aspiring technocrats see with AI is that it grants the rabble cognitive superpowers.

More specifically, *open-source* AI does.

Frontier models are proprietary. They have guardrails, training biases and actively scan and filter your conversations for potential safety issues. It’s arguably their purview to do so.

But it’s the open-source software movement that is the unfiltered Promethean agent here and just you wait – we’ll soon see what the ultimate target of this AI Doom cult really is: *open-source* LLMs.

*That* is why the frontier labs are gleefully courting government oversight – not because their own products are “too dangerous” (if so, then why are you going public?) but because oversight instantly creates an AI oligopoly – a cartel, namely Anthropic, OpenAI, and xAI, on the inside – and everything else, effectively *outlawed*  – at least that would be the preferred outcome.

This is the exact playbook I’ve told you about in the past, one that was described to me personally by an EIR from a Bay St. private equity firm in a meeting with two out of the four Big Four Canadian banks, back around 2015 – talking about Bitcoin:

“You get funded, you get out front, then you work with the government to create the regulations around it, and \*BOOM\*, you get to turn around *and pull the ladder up behind you”*

I still remember the look of absolute glee on his face when he got to that last bit. When I got home, I took a shower.

**What To Do About It**
-----------------------

I’ve been dutifully building out local compute capacity – both at home and thanks to our CTO, easyDNS now has a modest GPU farm out at our main datacenter.

I make it a point to download the latest open-weight models, Qwen, GLM, DeepSeek – which are, ironically, all Chinese…

![](https://bombthrower.com/wp-content/uploads/2026/09/VPN-openweight-meme.png)

But that won’t last. In fact, it’s already changing. We recently had “another country heard from” with the Japanese NII LLM-jp-4, Swallow, PLaMo and Rakuten AI 3.0 models weighing in.

France has Mistral, even Canada (yes, Canada) has Cohere Command A+ (my SolomonAI demo was, in case anybody needed to be told, a joke):

> The newest AI superpower is… Canada 🇨🇦. [pic.twitter.com/xb7nVy796n](https://t.co/xb7nVy796n)
> 
> — Mark E. Jeftovic (@jeftovic) [June 7, 2026](https://x.com/jeftovic/status/2063640130544353296?ref_src=twsrc%5Etfw)

And from elsewhere in “Big Tech,” there are Google’s Gemma 4.0 and IBM Granite.

Any of these models could be used to start iterating software, weights, and the next generation of more powerful open-source LLMs.

In other words, it’s over.

The incumbent ruling caste basically sees this, and viscerally realize that the entire edifice of Industrial Age institutions are ripe for existential disruption. I’ve written about this many times. We’re transitioning from a linear, top-down world to a network-shaped, decentralized one.

So the next best thing they can do, since they can’t put the genie back in the bottle, is to convince the masses that AI is bad, and rogue AIs are going to *literally exterminate humanity.*

**Hence the AI doom psyop** (with [another wave breaking out right now](https://www.yahoo.com/news/videos/top-ai-companies-investigating-tens-114712527.html) around swarms of rogue AIs), and, of course, this has become a political issue.

Now, the same people who were wearing N95 masks in 2025, adding their pronouns to their names, putting the Ukrainian flag in their bio, keying Teslas and stomping around in “No Kings” parades are now protesting data centers.

They want you scared, stupid and subservient, instead of spending your time creating software, building out [your personal business empires](https://theSovereignCapitalist.io), not to mention *analyzing public data* and proofreading all these hypothetical models they use as a justification for incessantly changing the rules.

What they really need are legions of frightened and stupid voters, not cognitively supercharged and wildly entrepreneurial or creative cyberpunks.

AI Doom vs AI Boom is a choice, and the choice is yours.

*Today’s post was a sneak preview from the forthcoming edition [of The Sovereign Capitalist](https://thesovereigncapitalist.io). You can get our latest [special report for paid members](https://thesovereigncapitalist.io/near-report-bonus/), and the next edition of the newsletter when you [grab this deal.](https://thesovereigncapitalist.io/near-report-bonus/)*