In 2022, a Google engineer named Blake Lemoine claimed his chatbot had gained sentience. At the time, I was running an AI games studio at a startup.
As someone who was trying to get early large language models (LLMs) to do much of anything useful at all, I found the story to be astonishingly naive and unhelpful. It absurdly anthropomorphized an experimental technology based on a total misunderstanding of the output.
Google did the right thing and firmly told the world that Lemoine had not in fact discovered a trapped child in the matrix, and fired him. But the meme that there are hidden entities inside the machine, a myth that goes back centuries, was not rejected by companies like OpenAI and Anthropic, they embraced it, and built on it.
In February 2026, an AI investor used a chatbot to write a spooky post on X about AI progress called “Something Big Is Happening” that got well over 100M views. As I wrote at the time, there was nothing substantial in the post, but it did illuminate what looks less like a technology and more like a fundamentalist religion in which Artificial General Intelligence (AGI) is the messiah, and the “Singularity” is the equivalent of the End Times.
On Tuesday evening, software engineer Jacob Coxon who had worked at OpenAI for three years, and moved to a job at Anthropic in July, posted a thread which asserted that he was quitting after roughly two months because “the people building AI earnestly believe that it could kill us all by the end of the decade.” This portent of doom was followed within minutes by a quote-tweet from his former colleague—who is the “Alignment Science Lead” at Anthropic.
Not only did Evan Hubinger agree with Jacob Coxon, he assigned a percentage to the possibility that “AI could kill all humans!” within a decade: more than 10%.
As of this writing, Coxon’s post has 158 million views. Within a day, he had done a substantial number of television interviews, including this deliberately alarming rant comparing AI to nuclear weapons on FOX.
“Yeah, so unfortunately we know how to control nuclear weapons pretty much. We don’t yet know how to control AI. 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 activity has been in the past.“
The hysteria generated from Coxon’s performances has been bipartisan and pervasive across social media. A number of Democrats have taken it at face value, including Bernie Sanders, who is being advised by people in the “Rationalist” “effective altruism” (EA) movement—an elite philanthro-capitalist ideology that serves to justify collecting massive amounts of wealth.
The EA movement and its overlapping x-risk institutions have been heavily financed by Silicon Valley billionaires: Musk funded Max Tegmark’s Future of Life Institute, while Peter Thiel was an early backer of Eliezer Yudkowsky’s Singularity Institute. It was also extensively promoted by now-imprisoned felon Sam Bankman-Fried, whose FTX invested $500 million in Anthropic before folding due to fraud.
Here is an excellent summary (gift article) of this movement and how deeply it connects to the AI industry. It is a complex story involving the transhumanist philosophy of Extropianism and the Peter Thiel-backed apocalyptic mythology of the Singularity, but it makes a good case that the chatbot industrial complex is more like a cult than a technology industry.
Simultaneous to the latest uproar over technological extinction, Anthropic released a whitepaper called “Economic Scenarios for Transformative AI∗” which puts forth three scenarios for the effect of the chatbot industry on the economy by 2030. Unsurprisingly, it puts forward no negative scenarios for growth.
In the “Modest” case, AI is as transformative as the internet. In the “Extreme Case” knowledge workers face drastic unemployment—but somehow GDP growth goes above 15%.
The financial model Anthropic used to make these predictions apparently did not factor in the more than 10% chance “AI will kill all humans!” that its own Alignment Science Lead predicted.
Put together, according to the chatbot industry, there are two possible futures:
1) AI becomes superintelligent and kills all humans
2) AI becomes superintelligent and replaces a large percentage of human knowledge work
As Nobel Prize-winning doomer icon Geoffrey Hinton said: “It will be very good or very bad.”
The first two scenarios appear to be opposites, but from a marketing perspective they sell exactly the same story: these companies are on the verge of creating something unimaginably powerful.
But both of these extreme futures make the assumption that AGI or ASI (Artificial Superintelligence) is near—and both promote technology created by OpenAI and Anthropic as being so powerful they need to be regulated by the government.
But there is a third possibility:
3) Chatbots reach diminishing returns. Scaling bigger becomes unsustainable and unprofitable. Demand for massive data centers never materializes.
If OpenAI and Anthropic cannot convince the public and the financial sector that scenario 1) or 2) is coming, and do not both successfully IPO, the entire economy will suffer a serious disruption. And if reality turns out to be scenario 3), rendering current investments in “AI” essentially a dead end, we have already witnessed the biggest misallocation of resources in human history, and it’s just getting bigger.
At least one-third of the S&P 500 is concentrated in stocks that require a massive growth story in AI. If that story falls apart, the ripples will be enormous.
The truth is that LLMs are a useful software feature that will be harnessed, commoditized and absorbed into the tech stack. That’s what’s actually happening all around us. It is a useful technology for compressing and querying large sets of data through natural language. It is not an oracle or a savior.
Sheer scale is not the answer to further developing artificial intelligence. The circular trillions being spent on gigawatt data center construction is mass psychosis.
The chatbot industrial complex is another even bigger version of the crypto industry—a moderately useful technology (blockchain) which was turned into an abusive Ponzi scheme through deceptive psychological operations by an antidemocratic minority of billionaires and techno-fascists.
The rotating science fiction stories that chatbots are either going to kill us, or make it so we never have to work again, are a pattern favored by abusers who allow for no middle ground in a relationship. The public is being oscillated between salvation and catastrophe, with ordinary outcomes excluded from consideration.
The reality is both much more boring and much more serious.
In my opinion, there is zero reason to believe AI will end humanity by 2030, or ever for that matter, but there is every reason to believe the industry that promotes it has inflated the largest speculative bubble in history based on science fiction stories.
Reality bites.
Thanks again to all who have sent messages, subscribed and donated recently in light of the judge’s ruling in the Flynn case. It is deeply appreciated.
NOTE: My articles are extensively researched and written by a human—me. No chatbots touched the words on this page.
Here is my DonorBox. 💙
Here is a GoFundMe for legal fees.🙏
It’s not hard to imagine how AI could cause massive disruption in many industries, even life threatening in sensitive places like hospitals or research labs. But, aren’t these applications dependent on electricity sources of some type? Seems that a solution for some pending ongoing threat from an AI unit would be to simply deprive it of power. What am I missing?
At my son's UCF graduation in May, Peter Lee from Microsoft Science was the commencement speaker. In his speech, he gave the typical pitch that "either it could usher us into a new and enlightened age for mankind, or, it could destroy us all". At no point that I recall did he advocate for passing strict regulations and laws to prevent the supposed doomsday scenario from taking place. In fact, aside from the typical "we need to work to make sure it's used for good" spiel, there wasn't any real elaboration on what we should be doing to prevent that alleged apocalypse.
After the speech, I stated to my family that anybody in an industry who would casually state that the industry that they work for might very well end all life on Earth, and yet still keep working for it instead of devoting every minute of their life to stopping that threat, is either a liar or an asshole...