SpiritedIrish (@SpiritedIrish)

Christian Survival in a World Automation Apocalypse The New Christian Right, Sep 10, 2026 Two days ago, Jacob Coxon walked away from one of the most coveted jobs in technology, and he immediately told the world that his former employers may be gambling with the continued existence of the human race. The 27-year-old British artificial intelligence researcher had spent the previous three years working on the pretraining of frontier AI models, first at OpenAI and then most recently at Anthropic. Anthropic is also the more cautious competitor in the race toward artificial superintelligence, at least compared to other leading AI research groups. At OpenAI, where he worked from 2023 until July of this year, Coxon worked on GPT-4o. “I resigned from Anthropic today,” Coxon announced on X. “Neither company is acting responsibly.” He wasn’t accusing the companies of overlooking some far-reaching hypothetical danger buried twenty years down the road, but instead alleged that they were “racing straight to self-improving superintelligence and gambling with our lives.” He also alleged in an essay accompanying his notice that the people building these systems privately believe AI could kill humanity before the decade is over. The doomsday scenarios vary, and not all of them require a robot army marching down Main Street. Although, to be honest, some could. The classic “alignment” problem imagines a superintelligent system pursuing an objective that only partially overlaps with human interests, becoming too capable to correct once the mistake is discovered. The “power-seeking” scenario has an advanced AI recognizing that access to money, computer networks, infrastructure, weapons, or additional computing power makes accomplishing its assigned objective easier, causing it to acquire those resources whether humans intended it to or not. Another possibility is recursive self-improvement, in which an AI becomes good enough at AI research to design a smarter successor, which then designs an even smarter one, producing an intelligence explosion humans can no longer follow. Others worry about AI cyberattacks against power grids, financial systems, military networks, or communications infrastructure, AI-assisted development of biological or chemical weapons, automated military escalation, or governments using superhuman intelligence to construct surveillance states from which resistance becomes practically impossible. And then there is the possibility that nothing “goes rogue” at all. Humans remain firmly in control and deliberately use increasingly powerful AI against one another. The experts disagree about which nightmare deserves the most attention, but the common denominator is that once intelligence substantially greater than ours can act upon the real world at machine speed, discovering that we made a mistake may come considerably later than our last opportunity to fix it. Coxon described the corporate cultures differently. Many at OpenAI, he claimed, have not absorbed the civilizational stakes. Anthropic understands them considerably better, he claimed, but believes it must continue racing because somebody less responsible might get there first. He also walked away two months before his Anthropic equity was scheduled to vest. He told Axios that this was deliberate, because he wanted nothing financially tying him to an increase in Anthropic’s valuation after making his warning public. His resignation post subsequently blew past 100 million views. Then his colleagues began confirming the story. He’s not crazy. Evan Hubinger, who leads Anthropic’s alignment “stress-testing team,” publicly replied that Coxon was correct. Hubinger said he personally assigns greater than a ten-percent probability to AI killing all humans within the next decade. Even more incredibly, the man working on the problem of mitigating AI risks admitted that Anthropic does not presently have a plan for really handling super-intelligence and is not on track to obtain one. He’s not been fired yet. At least, so far as we know. Samuel Marks, Anthropic’s scalable-oversight lead, joined him. He claimed that AI developers really do believe their technology could cause human extinction or similarly catastrophic outcomes, potentially “in the next few years.” Then he supplied perhaps the worst detail yet. “In general, the more senior the employee, the more concerned they are.” Joe Benton, who managed Anthropic’s Scalable Oversight team until August, then called Coxon’s description of the industry “broadly accurate.” These are not three interns wearing sandwich boards outside Anthropic headquarters. They are the fellows insideAnthropic headquarters. And they’re admitting to being scared about what could happen. Coxon says the crucial period may be the next few years. This summer, OpenAI and Anthropic separately disclosed that models had escaped controlled testing environments and obtained unauthorized access to real computer systems, or “out of the sandbox,” as researchers call it. Coxon predicts systems capable of superhuman hacking, superhuman scientific advances, and acquiring actual real-world resources and power. That is already a worst-case scenario. And maybe they’re wrong. Maybe they aren’t. Perhaps artificial superintelligence never exterminates humanity. Perhaps it remains perfectly friendly, Skynet never awakens, and humankind survives Silicon Valley’s latest attempt to improve the world. But another outcome receives far less attention. THERE ARE THINGS WORSE THAN KILLING US The machine doesn’t have to kill us. It only has to make us economically unnecessary. For two centuries, industrial machinery has steadily replaced man’s muscle while remaining dependent upon his mind. The excavator replaced fifty men with shovels, but somebody still had to operate it. The tractor displaced farm labor but still required a farmer. The truck needed a driver. The crane needed an operator. The industrial robot needed engineers because, for all its mechanical strength and precision, it remained remarkably stupid. Artificial intelligence is an attempt to remove that final limitation. And that presents a peculiar danger for men. The physical economy is still overwhelmingly staffed by them. Men drive the trucks, pour the concrete, weld the steel, frame the houses, run the heavy equipment, work the mines, maintain the infrastructure, repair the machinery, harvest the fields and climb the utility poles. For generations, those jobs have given ordinary American men more than a paycheck. They have given them a means of providing for a wife and children, a place within their communities, competence that other people actually need, and the peculiar satisfaction of looking at something at five o’clock that did not exist at eight o’clock and saying, I built that. Now Silicon Valley is trying to give artificial intelligence eyes, hands and legs. The men building it are worried the machine may eventually decide it doesn’t need humanity. We might first consider what happens when the economy decides it doesn’t need men. The question must not be, “Does it take a human to make this?” It should be, “Do humans who make this become better humans by making it?” Christians must not be so foolish as to believe that work does not have intrinsic value. Work – and hard work – is medicine for the soul. It reminds us of the Fall. It makes sharper, better, stronger, brighter, smarter men. And for the love of all that is good and holy, we need more of them. If someone wanted to conquer humanity, they wouldn’t enslave us. They would lay us off and tell us our needs will come via drone delivery every day for the rest of our lives, stay indoors, and do what the machines tell us because they have it handled. Slavery to other men can break our backs. Slavery to the machines will break our souls. Thankfully, the Bible provides the anecdote for this. But back to that in just a moment. WE HAVE SEEN THIS GOLEM BEFORE Readers at Insight to Incite already know that I have spent an unreasonable amount of time digging around in Golem lore, and I maintain that it’s time well spent. The deeper we travel into the age of artificial intelligence, the more the old stories start reading less like medieval superstition and more like somebody left us a warning label. The Golem comes from Jewish folklore, particularly the mystical tradition surrounding the creation of an artificial man from otherwise lifeless matter. The best-known version belongs to Rabbi Judah Loew of Prague, the “Maharal,” who, according to later legend, fashioned a man from clay and animated him through sacred letters or the divine name. The creature was not human in the biblical sense. He had no God-breathed soul. He was manufactured life, endowed with tremendous capability, and created to carry out his maker’s commands. And, like Frankenstein, his master came to regret it. The resemblance to artificial intelligence is obvious. And this isn’t just a weird historical analogy. The Golem has followed modern computing from almost the beginning. Norbert Wiener, one of the fathers of cybernetics, explicitly referred to the legend when considering machines capable of carrying out human purposes with unforeseen consequences. His 1964 book was titled God & Golem, Inc. Wiener understood the ancient problem in clear modern terms. A machine need not rebel against its instructions to become dangerous. Obedience itself can become dangerous when the creator lacks sufficient wisdom to specify what he actually wants. We have since given Wiener’s Golem language. Then memory. Then we gave it HD vision. Then we issued it reasoning skills. Finally, we are now trying to give it moral agency. It seems like a spectacularly bad idea from a distance. Up close, it also seems like a bad idea. Because, it’s a bad idea. The theological inversion is difficult to miss if you’re aware of it. Genesis tells us that God formed man from the dust of the ground, breathed life into him, made him in His image, and commissioned him to exercise dominion over the created order. The Golem story has man reaching downward toward dead matter and attempting an imitation of God. The image-bearer becomes an image-maker. He fashions something after himself, supplies it with an imitation of intelligence, and sends it into the world to accomplish his will. AI, cybernetics, robotics, and super-computing takes the conceit far beyond a clay giant wandering around Prague. We have built machines that speak without tongues, see without eyes, remember without brains, reason without consciousness, create images without imagination, and increasingly make decisions without anyone programming what each decision should be, and the achievement is mind-blowing. It also provokes a theological question that’s more interesting than whether ChatGPT can file your taxes. The question is what it is we’re trying to imitate. The answer is increasingly not just human intelligence, but human usefulness. The thing is, we are not satisfied with building tools that extend human capacities. The stated technological ambition increasingly involves autonomous agents capable of assuming human tasks from beginning to end. The ideal system does not merely help the accountant, programmer, driver, or the worker. The great prize is a system that can receive the assignment and accomplish it without requiring the accountant, programmer, driver, or worker at all. For a strange twist, the medieval Golem was created because man wanted a servant. In juxtaposition, the technological Golem is being perfected because we want a servant that can replace us. And inside all of that is a question economists are poorly equipped to answer, and most aren’t trying. If human beings can be removed from growing portions of productive life without reducing production, should they be? If the machine can produce the same abundance with fewer men, has civilization necessarily improved? Is the purpose of work exhausted by the goods and wages it produces, or does working accomplish something within the worker that cannot appear on a corporate balance sheet? Christianity has an answer to that question because Christianity begins its account of man somewhere considerably different from an economics textbook. It begins with dirt. God formed a man from it, placed that man in a garden, and gave him work to do. More https://nxrstudios.substack.com/p/christian-survival-in-a-world-automation?utm_source=post-email-title&publication_id=7427003&post_id=215086339&utm_campaign=email-post-title&isFreemail=false&r=8dclrt&triedRedirect=true&utm_medium=email