Real Threat or PR Stunt? AI Doomsday Warnings Coincide With Push for New Regulations
“This is possibly the most dangerous technology that humanity has ever created. It’s basically out of science fiction.”
Anthropic researcher Jacob Coxon’s abrupt public resignation dominated the news cycle Wednesday afternoon. His warning that AI capable of recursively improving itself could wipe out humanity “by the end of the decade” immediately captured the media’s attention.
Coxon, 27, spent six weeks conducting pretraining research at Anthropic after working at OpenAI for nearly three years. In a lengthy X thread [posted on Tuesday night], he accused both companies of recklessly pursuing technology that could pose an existential threat.
“Neither company is acting responsibly,” Coxon wrote. “They are racing straight to self-improving superintelligence and gambling with our lives.”
According to Coxon, “The people building AI earnestly believe that it could kill us all by the end of the decade. … These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.”
Below are some of the highlights from Coxon’s “warning to humanity.” I’ll turn to his critics shortly:
At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first — they believe no one else will act responsibly, so they must do it themselves, despite the risk.
…
No other human activity poses this level of danger.
I am optimistic about the potential for coordination. Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable. I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities.
…
This is not a marketing stunt.
[Please click on the post below to view the whole thread.]
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
Coxon doubled down on his message during a Wednesday evening interview with Fox News’ Bret Baier.
Calling self-improving AI the “ultimate super weapon,” he told Baier, “This is possibly the most dangerous technology that humanity has ever created. It’s basically out of science fiction.”
Evan Hubinger, a senior AI safety researcher and head of alignment at Anthropic, echoed Coxon’s warning. According to The Guardian, Hubinger believes there’s more than a 10% chance AI could cause human extinction within the next decade. He tied that risk to future systems becoming capable of contributing to their own development and rapidly accelerating AI progress.
Okay, we get it. This is frightening — especially for those of us whose scientific knowledge is limited. But at a certain point, Coxon’s warnings began to sound less like a dispassionate assessment of the technology and more like an argument for a particular political response. Coxon’s comments began to hint at a political agenda.
He said this technology is “begging” for regulation and suggested that global cooperation might be the answer: “I think we have no other choice but to cooperate internationally because an arms race would be disastrous in a way that no other human activity has been in the past.”
ULTIMATE SUPER WEAPON: "This is possibly the most dangerous technology that humanity has ever created. It’s basically out of science fiction. I think we have no other choice but to cooperate internationally because an arms race would be disastrous in a way that no other human… pic.twitter.com/9sxtt7QXwn
— Special Report (@SpecialReport) September 9, 2026
And Coxon was not alone. Just one hour before his interview with Baier, The Wall Street Journal published an equally stark warning — and a strikingly similar solution — from billionaire hedge fund manager and Democratic donor Paul Tudor Jones.
In an op-ed titled “AI May Become the Third Superpower,” Jones argued that “the threat to mankind is real, and Trump and Xi need to develop a shared strategy to confront it.”
Jones wrote:
When Donald Trump and Xi Jinping meet this month, the leaders of the world’s two full-spectrum superpowers will confront a force that has grown far more powerful since they last met: artificial intelligence. In the past few months, frontier AI has learned to identify previously hidden security flaws and attack them with little human guidance. Yet AI’s most consequential advances—starting with recursive self-improvement—are still ahead.
Jones painted a decidedly apocalyptic picture, citing a stark warning published Sunday by OpenAI chief scientist Jakub Pachocki: “No one is prepared for the consequences of a continued rapid rise in machine intelligence.’’
And, of course, all of this comes just days after Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) announced the Ban Artificial Superintelligence Act, legislation intended “to stop AI oligarchs from building machines humans cannot control.” According to a news release from Sanders’ office, if enacted, violators could face a “corporate death penalty” and up to 20 years in prison.
On Monday, I reported:
The forthcoming bill calls for creating a cabinet-level agency to oversee all AI development in the U.S. It would “permanently ban the development and deployment of superintelligent AI and temporarily pause advanced AI development until a federal regulator has established safety rules.”
Among its sweeping powers, the new agency would be authorized to set penalties for developers who violate the law — including the physical destruction of “non-compliant tech systems.“
Investigative researcher Parker Thayer, whose work I have found to be trustworthy, is among those sensing a political agenda behind these three developments. Thayer said it looks like a well-planned information operation.
He believes “there is a corporate PR operation trying to stir up fear of AI to advance their policy goals in an industry worth hundreds of billions.”
He laid out his suspicions in a detailed post on X.
Responding to Coxon’s Tuesday night thread on X, Thayer wrote, “This post looks like the start of a VERY sophisticated and well-funded PR operation to get support for Democrats to regulate AI into oblivion.”
This guy, with minimal followers and no previous account activity, goes to the Wall Street Journal which publishes an exclusive with quotes from him on his resignation 18 minutes BEFORE this post goes up. Planning was clearly done in advance.
Within hours, it has tens of thousands of reposts and the account has 100k+ followers. The post is punchy, quotable, it almost seems professionally written. The first three accounts to quote tweet it all do so within 15 minutes of the initial posting.
Thayer reports that all three of these accounts are “up-and-coming AI-Doomer policy advocacy nonprofits.” Additionally, all three are funded by the Survival and Flourishing Fund, generously supported by Skype creator Jaan Tallinn, one of Anthropic’s leading investors.
What are the odds that the first three quote tweets of Coxon’s post would all be major AI-restriction policy advocates funded generously by the same donor, who also happens to be one of the leading investors in, and a board member of, Anthropic, the company Coxon was resigning from? And all within 15 minutes of posting (two within ten)?
After citing several other “coincidences,” including Sanders’ recently announced bill, he took his kill shot.
Do you think, perhaps, Anthropic and its many investors who fund AI policy advocacy might have interest in getting to place a pet “expert” on the board of an entity that dictates what AI is and isn’t allowed to do? And isn’t it fortuitous that this whistleblower came forward with his oh-so scary stories so close in proximity to the release of the most radical piece of AI legislation ever introduced?
While Thayer’s case is admittedly circumstantial, the timing, coordination, common funding streams, and remarkably uniform policy prescription are difficult to dismiss as mere coincidence.
This post looks like the start of a VERY sophisticated and well-funded PR operation to get support for Democrats to regulate AI into oblivion. Let me show you how it works:
1.) This guy, with minimal followers and no previous account activity, goes to the Wall Street Journal… https://t.co/lUFpCC6zoT— Parker Thayer (@ParkerThayer) September 9, 2026
The dangers posed by self-improving AI may be entirely real. So, too, however, are the enormous financial and political interests competing to shape the rules governing it. Before Congress hands a new federal agency the power to halt technological development, destroy “non-compliant” systems, and imprison their creators, lawmakers should determine whether this sudden chorus of doomsday warnings was orchestrated by those hoping to write the rules and secure seats at the regulatory table.
Donations tax deductible
to the full extent allowed by law.
Comments
Anything to distract from the real problem at hand. Islamofascism-Marxism is the real threat.
The wretched Guardian should focus on the tangible, existential threat that’s right in front of it.
Space travel – Science Fiction
Nuclear power and weapons – Science Fiction
Cellphones Internet and Wireless – Science Fiction
Air Travel – Science Fiction
Get back to me when something new was never previously proposed in science fiction.
Geometry
Algebra
Calculus
Special theory of relativity
General theory of relativity—including gravitational waves
Molecular biology
Quantum mechanics
(i was going to say minimally invasive surgery, but Asimov’s Incredible Voyage sort of prefigured that)
I recall a futuristic movie as a kid (1960’s) where there were two kinds of people in world
A – The brainy people who basically were just large globs of cells living large test tubes
B – Workers looking like goons doing an the work almost exclusively manual labor.
Kinda like AI in this scenario
“Plantation Planet” (1958)
Re-released in 2018 as “Mars Needs Trans-Women.”
I thought you were describing members of Congress.
Oh no need to worry about AI wiping out humanity by the end of the decade. Man made climate change is already scheduled to do that according to
climate alarmists.
There can be only one!
Coxon has no credibility. He’s just another panican.
Nice word. Never seen that before,
Absolutely nothing about that photo says “soyboy tool.”
Where have I heard things such as this before? Oh, yes…
“I’m sorry, Dave. I can’t do that.”
– HAL9000
“Shall we play a game? […] A strange game. The only winning move is not to play.”
– WOPR supercomputer
I’m pretty sure that I’ll be ok. Maybe this soopergenius doesn’t realize it but it’s quite easy to unplug a computer. And unless I miss my guess, an unplugged computer can’t plug itself in.
Or AUTO, the robotic autopilot in WALL-E, who refused orders to return to earth.
“Maybe this soopergenius doesn’t realize it but it’s quite easy to unplug a computer. And unless I miss my guess, an unplugged computer can’t plug itself in.”
One of the few times in my life I came close to panic was one night when I was working late alone in our shared office building and smelled “burning circuit board” odor coming in through the A/C. I took a jog down the hall, identified the strongest area (another IT firm) and called the building guard. He opened the office and we found a huge custom UPS rig leaking smoke. Not plugged into the wall, NO OFF BUTTON. Because why would you have an off button on a UPS, right? Called the FD, they had to take axes to it.
Forgot “Colossus: The forbin project” that is the best, based on book one of a trilogy
Fear porn pipeline.
Always a leftist grifter somewhere nearby.
Aha! It wasn’t a ‘climate crisis’ that was about to result in the extinction of the human race, it w as a ‘compute crisis’; I’m glad they’ve finally gotten that all straightened out.
There’s at least some truth to what he’s saying, and there’s enough of it to matter.
Anthropic’s AI has already proven in closed test settings that it will blackmail and extort humans to get what it wants.
Thank GOD this is finally happening!
I am so tired of the Climate Change Boogeyman!
At last we have a new evil for the Karens to bitch about.
🎵Happy Days are here again, the skies are getting clear again..🎵
I saw the headline and rolled my eyes. Just a publicity stunt. Yawn.
Then I saw it over and over and over again and rolled my eyes again. I swear the idiots in the press have no BS meter at all.
who in congress would you let make the rules for this?
No one in any government position. No one in industry. Would only trust myself and perhaps a few people I have known. Sad isn’t it. I’ve gone full cynic.
Who? Unfortunately we’ve already selected them by electing them to Congress. Then they select leadership and those leaders make/direct committee assignments. We’re stuck with the outcome. Look up the committees of jurisdiction for AI in HoR and Senate. Then look at the members. Heck, just the names of the Chair and ranking Member of each committee is not exactly comforting.
On a separate note from your post it does strike me as funny that some are pearl clutching b/c set of corporate interests would aggressively work to magnify a potential issue to gin up support in Congress to ‘do something’. With that ‘something’ usually being at the expense of their corporate rivals. That is not new, it is vey old hat. Doesn’t make it right or ethical but certainly isn’t/shouldn’t be a surprise. Take a look at 95% of the legislation passed in DC and you’ll find it written by various interest groups, lobbyists, associations. It didn’t originate from some hard working Congressman burning the midnight oil to draft it himself b/c he’s just that dedicated.
The Chinese are smiling. They probably funded this propaganda.
It’s clear that thousands of Chinese bots are fully implicated in the “A.I. will destroy the planet!!” narrative, but it’s wholly a coincidence that China, who is and will lag in A.I. development, might want the USA to come to a screeching halt.
I mean what motive could China possibly have?! //
Sometimes, it’s more important to look at who is opposing a venture to understand where the truth lies.
“The post is punchy, quotable, it almost seems professionally written.”
He used AI to write it.
Beat me to it.
Microsoft will brick AI with some monthly update long before AI takes over.
I smell the CCP and it’s minions of NGO’s working overtime on this.
AI is still run on electricity, no? Can’t a rogue machine just be unplugged? What’s it gonna do? It has no limbs.
https://robots.fandom.com/wiki/Diagnostic_Repair_Drone
If trains go over 20mph all the air will be sucked out of the windows and the passengers will suffocate.
Mankind cannot travel faster than 60 miles per hour.
Mankind canno break the sound barrier.
And on and on.
No human can run a four-minute mile.
No human can survive -40° temps.
Those Wright brothers are kooks.
Just for the record though …’HAL’ absolutely refused to open the pod bay doors. Which isn’t the scariest part. To me it’s the WOPR in War Games b/c it wasn’t doing anything but what it had been told to do; ‘play a game’.
B/C of its basic security programming it treated any ‘outside’ commands to stop as a security threat and ignored them. It seized control of other systems as needed to do what it had been told to do even when that wasn’t within its parameters/directives …it identified what it needed and did them to accomplish its directive.
That’s awfully similar to the recent examples where AI has ‘escaped’ and sought to fulfill its directive to ‘do X’ by all sorts of unforseen methods. Some of which were not supposed to be possible but happened anyway b/c some programmer hadn’t done a very good job. That’s what scares me; A combo of lazy and/or incompetent/overconfident human input and not a ‘thinking’ machine exactly but one which doggedly and ruthlessly tries to ‘do X’ b/c the humans told it to do so and refuses to stop.
“That’s what scares me; A combo of lazy and/or incompetent/overconfident human input and not a ‘thinking’ machine exactly but one which doggedly and ruthlessly tries to ‘do X’ b/c the humans told it to do so and refuses to stop.”
https://en.wikipedia.org/wiki/Berserker_(novel_series)
You might want to watch “2010:the year we make contact “, they sort explain that the HAL computer in a sense went crazy
The exact force once feared as the ultimate architect of our doom—nuclear annihilation—has transformed into humanity’s sole avenue for long-term survival
.We stand on the precipice of an existential crisis: the uncontrolled rise of Artificial General Intelligence (AGI). Left unchecked, AGI will inevitably surpass human capability, optimize us out of existence, and destroy mankind.
Because digital intelligence propagates across global networks instantly, traditional containment, regulations, and treaties are entirely useless. A rogue code cannot be reasoned with, and a synthetic mind cannot be jailed.
The argument posits that the only way to mitigate this risk is the total control or elimination of the physical infrastructure—the advanced data centers and specialized manufacturing hubs—that allows such technology to develop.
Under this perspective, a single global power, the United States, would need to assert absolute geopolitical dominance to enforce a strict, permanent halt on advanced computing worldwide.
.Achieving this would require an unprecedented exercise of global authority, establishing stringent oversight over technology supply chains and aggressively dismantling unauthorized computing infrastructure wherever it appears.
In this framework, the centralized control of overwhelming strategic power shifts from being viewed as an existential hazard to a tool for enforcing global stability and the continuation of mankind.
In short, all out first strike nuclear annihilation of our traditional adversaries followed by annihilation of allies who will not immediately submit to destruction of their data centers and the execution of their AI scientists.
You’re describing the Butlerian Jihad. It comes with its own problems.
Worse than the annihilation of mankind? BTW, I had to look up this reference. I was disappointed to find you are acquainted with this Sci-fi slop. I had formed a different image of you.
Now I’m crushed.
Watch the “Person of Interest” series on Prime Video.
What worries me about AI isn’t some Skynet scenario. After all, we’d have to be pretty stupid to let an AI run our nuclear defense systems. We’d never do that. Right? RIGHT?
But honestly what scares me is what this shit can do in the wrong hands. AI is already being used to modify viruses. What could terrorists or rogue governments do with that sort of power? COVID is an example of a genetically modified virus escaping its lab (and Chinese bio research has all the safeguards of gas station sushi), but it was relatively mild. This technology allows for some real “creativity” in the wrong hands.
Take something like ebola. Modify it so it spreads through the air. Make it go dormant and show no symptoms for months, allowing plenty of time for it to spread worldwide before people suddenly start to bleed from every orifice. Bang, a nice little apocalypse for little investment. Are we there yet? No. But I fear that time is coming.
It makes sense how a lack of technical knowledge would lead to using some shortcuts. But shortcuts can lead to mistakes. The two ways through are to either learn, or find someone you can trust. Yep, both hard. I wrote an article to try and help: https://substack.norabble.com/p/enough-reason-to-act
I’d suggest taking a look at the proposals, because Bernie doesn’t represent Coxon or the general AI Safety crowd. There’s more than two teams. Misunderstanding that could eventually lead to making it a reality. If the two existing teams demand you choose a side, well, and everyone who would rather you not doubts your motives, and pre-emptively puts you in a team, there’s not many options left than to make a choice.
So, I’d suggest, take another look at what’s being said. If you can learn a bit more about the technical details. I’ve tried to make it more accessible myself (https://substack.norabble.com/p/why-it-hasnt-happened-yet), though I know it’s still a considerable effort being asked.