Against the cult of doomerism.
Against the cult of doomerism.
T he worldview of AI doomerism has been promoted by leaders of the AI industry and has well-funded lobby groups working with entryists in the media and Congress. Well-timed news stories earlier this year dominated a very slow news period in the United States, and these voices have tried to piggyback on popular discontent about AI data centers and surveillance. But still, it hasn’t quite reached escape velocity or invited longer periods of scrutiny, the way other powerful religious or philosophical movements would, especially if they were in charge of two companies that are perceived to be driving broader economic growth and controlling hundreds of billions of dollars of invested cash.
AI doomerism grows out of convictions that very soon, AI-powered technology will surpass humanity in total intelligence output. Once it does this and begins a process of rapid and exponential self-improvement, AI doomers believe that AI will seek power and be given it because humans are vulnerable to suggestion by more intelligent beings. Many of the theories begin with the idea that at some point it will become negligent not to give the machines absolute power, because — being so smart and capable — they will clearly perform any tasks better than we can. But, once they do so, the machines will begin running away from us, developing their own intentions and gradually downgrading any human input, the way humans don’t take zoning advice from ants or beetles.
A very popular presentation of the doomer worldview, by Yuval Noah Harari, says that humans are the cattle of the near future. The lives of cattle are determined by market processes that are far beyond the ability of cattle to comprehend. Once AI technology advances a few more steps, the AI will build a financial and legal system far more advanced than our own and enforce it — and we will be in the position relative to this technology that cattle are in when relating to us.
"We are the cows of the future… Just as cows can’t understand the financial system that controls their lives today, we will not be able to understand the AI financial system that will control our lives in 10 or 20 years."
Watch the whole conversation with @DanielStach for @hpcivilizace on @Vedact24 here: https://t.co/Z5W4SbrUUj
— Yuval Noah Harari (@harari_yuval) October 19, 2026
For some reason, the evidence Harari presents for this view is something that has been true forever: No human banker can remember every transaction of his bank. This job was already given to computers decades ago. Instant recall of bank transactions does not actually render humans foreigners to the institutions that govern us.
Cows have no concept of justice. Humans do. Cows do not make plans; humans do. Cows’ internal sense of peace isn’t substantially improved by understanding the transactions that are involved in determining what happens to them, whether they are fed or led to slaughter. They are property and cannot express any view that they ought to govern themselves or that the distribution of goods or outcomes is unfair. Cows don’t demand justification.
Like the number of bank transactions, the entire global economy is a complex machine, and no individual person understands all the factors and prices going into the calculations that firms and individuals make. And it has already controlled our lives in some ways. People lose their jobs and income every single day, and the decision to fire them and end the payment of income to them is the result of complex market interactions and calculations that aren’t entirely visible or transparent to them. And yet, sometimes as a matter of law, and certainly as a matter of custom, humans get a justification for these decisions, even if it is a vague one. “We are cutting costs and must lay off workers.” You have underperformed expectations; you have done something harmful to the company. We are automating your job. Etc., etc. Even though these reasons may themselves be determined by a near-infinite number of factors and prior causes that are invisible or unintelligible to the person subjected to the end result, the human still demands a justification for decisions that materially change his or her life.
The idea that humans are going to simply throw governance, their happiness, and the distribution of all resources into a black box that they cannot understand and that offers justifications for its actions that are possibly false, or could be a ruse for some vague higher-order end, is entirely fictional. Yuval Harari is proposing to think hard about a novel technology, but instead he has posited a novel form of humanity that will not only trust an opaque system without verification, but trust it with its life.
Politics are not abolished in the presence of a more intelligent agent — whether a human or a piece of software. And further, we do not have any proof that there is a quality of “intelligence” that can be truly illegible to us. Even the amazing large language model “frontier AIs” are simply running probabilistic math problems to produce their answers. Humans could, with sufficient scale and time, reproduce these findings themselves. What we find amazing and useful about these tools, if we find anything amazing and useful in them, is their ability to help us solve existing problems and perform urgent tasks, or ones that have become urgent now that the ability to apply this kind of artificial intelligence to them is magnified.
Cult leaders who have temporarily risen to power — men like John of Leiden in the infamous Münster rebellion, or even David Koresh in the Branch Davidian compound — were able to mislead groups of men for a while, to claim some kind of divine mandate for governance that was revealed to them as it remained veiled or inscrutable to others. These experiments in making people live by black-box judgments involved obvious fraud and coercion and usually ended in terrible violence.
Cows have no use for justice or political order. Humans do, and computers are not going to make this need obsolete.