# Mustafa Suleyman Says 18 Months. I'm Taking My Time. Last Tuesday I sat with a man in his fifties who had just been told his data-entry job — the one he'd held for eleven years, the one that paid for his daughter's school uniform — was being automated out of existence by the end of the quarter. He wasn't angry, which was almost worse. He was just quiet, in the way people go quiet when the world has confirmed something they always suspected about their own dispensability. That same morning, Mustafa Suleyman, Microsoft's AI CEO, told a conference audience that human-level artificial intelligence is roughly eighteen months away. I keep trying to hold both of those facts in the same hand. I am writing this slowly. I want to say that up front, because the entire industry surrounding this question is built on speed, and pastors are, by trade and by calling, supposed to be people who can sit with someone for an hour and not check their phone. If you have come here for a hot take, I commend to you any of the 47,000 LinkedIn posts published yesterday on exactly this subject. What I want to do is something duller and, I hope, more useful: think out loud about what a Christian congregation in a tech city ought to be doing while Silicon Valley counts down its own arrival. ## What Suleyman Actually Said, and Why Pastors Should Read It Themselves Before anyone in church leadership forms an opinion about Suleyman's eighteen-month claim, they should read it. I mean the actual remarks — not the headline, not the YouTube clip with the alarmed thumbnail, not the tweet that ricocheted around your WhatsApp. Suleyman has been carefully calibrating his language for years. In his book *The Coming Wave* he speaks of the "containment problem" and of a series of "capability thresholds" that AI systems are now passing in close succession. The eighteen-month figure is less a date on a calendar than a rhetorical device meant to focus the minds of policymakers and shareholders. This matters, because the temptation in the church is to outsource our thinking to whichever Christian commentator we already trust. I do this myself. I am four podcasts deep into the AI question before I have read the primary source, and then I find that my opinion is really an averaging of other people's opinions — which is to say, no opinion at all. Augustine spent years insisting that pastors must be readers, not because reading makes us clever, but because the alternative is to be moved by every wind of doctrine and every viral clip. So: read Suleyman. Read Demis Hassabis. Read the doomers and read the boosters. You will not become an expert. You will become marginally less foolish, which is the most any of us should hope for on most topics. ## The Eschatology Already Baked Into the Forecast Once you read the primary sources, something becomes obvious. Silicon Valley has an eschatology. It does not advertise itself as one, because the people writing it would be embarrassed to be told they are doing theology, but it is unmistakably a theology of last things. There is a coming event. The event will inaugurate a new age. The new age will solve the problems of the old one — disease, scarcity, perhaps even death. Those who align themselves with the coming event will flourish; those who resist will be left behind. Replace a few nouns and you have a tract from a nineteenth-century millenarian sect. Augustine saw this coming, in a manner of speaking. In *The City of God* he distinguishes between the city of man, which is always building towers and always disappointed, and the city of God, which is patient because its hope is not contingent on the next innovation. "Two loves," he writes, "have made two cities: love of self, even to the contempt of God, an earthly city; and love of God, even to the contempt of self, a heavenly city." The point is not that technology is the earthly city — that would be too neat. The point is that any project organised around human self-transcendence eventually has to answer a theological question it pretended it wasn't asking. Nietzsche, who hated Christianity for being too soft, would have admired the boldness of the AGI promise. The Übermensch — the human who finally throws off slave morality and seizes godlike capacity — was Nietzsche's wager. What he could not have predicted is that we would try to build the Übermensch in a server farm rather than become him. Whether the silicon version is more flattering or more humiliating to the original aspiration, I leave to others. The pastoral point is simpler: when an industry tells you it is about to deliver salvation, the church's first job is not to celebrate or to panic. It is to ask, quietly, salvation from what, and salvation to what, and at whose expense. ## What We Get Wrong When We Talk About 'Human-Level' The phrase "human-level intelligence" is doing a tremendous amount of work in the AGI conversation, and almost none of it is honest. To say a machine has reached human-level intelligence, you must already have decided what human intelligence is *for*. And there, the conversation gets thin very quickly. If by human intelligence we mean the capacity to pass exams, generate plausible text, write working code, and produce a passable sonnet on demand, then we are nearly there — or perhaps already there, depending on your benchmark. But this is intelligence stripped of everything that has ever made human beings interesting to other human beings or, dare I say it, to God. It is intelligence as a productivity metric. The biblical category here is the *imago Dei*. Genesis 1 tells us human beings are made in the image of God, and the church has spent two thousand years arguing about exactly what that means — rationality, relationality, dominion, creativity, the capacity for moral love. What it has never meant is "the ability to complete benchmark tasks at or above the median human score." If we accept the AI industry's working definition of intelligence, we have already lost the more important argument before any machine arrives. The real danger is not that we build a machine that thinks like a human. The real danger is that we redefine what it means to be human in such a way that the machine qualifies. This is happening already, in subtle ways — in how schools measure children, in how managers measure workers, in how dating apps measure desirability. The AGI debate is the loud version of a quieter capitulation that has been going on for decades. Paul writes to the Corinthians: "Knowledge puffs up, but love builds up. Anyone who claims to know something does not yet know as he ought to know." That is not a sentence designed to win a debate at a TED conference. But it might be the sentence we need pinned above every server room. ## The Man With the Quiet Face, and What the Market Cannot Measure I want to return to the man I sat with on Tuesday, because abstractions cost nothing and his job did. He told me, over a cup of coffee that he insisted on paying for, that the worst part was not the financial fear, though that was real. The worst part was that nobody at his company had told him in person. He learned by email — an email that contained a paragraph about "the changing landscape of skills" and a link to a retraining portal. What does the church have to offer this man? Not a job, though we should help him find one. Not an opinion about Suleyman's eighteen months, though we may have one. What the church has is a story in which his dignity does not derive from his economic function. He is loved before he is useful, and he will be loved when he is no longer useful , which is the terror that every human being carries with them and which the labour market cannot answer. Micah 6:8 is one of those verses that gets quoted to the point of cliché, but cliché is just the dust that settles on a truth nobody has the energy to refute. "He has shown you, O mortal, what is good. And what does the Lord require of you? To act justly and to love mercy and to walk humbly with your God." Notice what is not in the list: to be productive, to keep up, to remain relevant, to pass the capability threshold. The Lord requires of this man what he requires of Suleyman, and the requirement is shockingly indifferent to the question of whose job is automatable. A church that has internalised this will offer the displaced worker something the retraining portal cannot. It will offer a community in which his Sunday is not measured by his Monday. This is not consolation prize theology. It is the actual order of things, finally made visible. ## Taleb Was Right: We Are Terrible at This Nassim Taleb spent much of the last twenty years pointing out that human beings are dreadful at forecasting, particularly the timing and magnitude of large discontinuous events. We construct narratives after the fact and then convince ourselves we predicted them. He calls this the narrative fallacy, and the AI conversation is drenched in it. Eighteen months. Why eighteen? Why not twenty-four, or thirty-six, or six? The honest answer is that eighteen months sounds urgent enough to motivate action but distant enough to remain unfalsifiable in the short term. By the time the eighteen months have elapsed, the goalposts will have been redefined, the benchmark will have shifted, and someone will produce a paper explaining why we either did or did not pass the threshold, depending on which conclusion is more useful to the writer. This is not a reason to dismiss the AGI question. It is a reason to slow down the panic and the hype with equal pressure. Taleb's point is that Black Swans , the genuinely transformative events , are by definition the ones nobody saw coming. The events we spend years predicting tend to disappoint on arrival, while the ones that actually reshape the world arrive on a Tuesday with no warning. The 2008 financial crisis was a Black Swan; the steady drumbeat of AGI predictions is not. So my counsel to pastors is this: take the technology seriously, take the displacement of workers seriously, take the cultural shift seriously. But do not take the timeline seriously. Anyone who tells you they know what 2027 looks like is selling you something , including, with respect, Mustafa Suleyman. ## The Particular Temptations of the Church in a Tech City I plant a church in a part of London where one in three people in our congregation works in or around the tech industry. We have engineers from DeepMind, product managers from companies whose names you would recognise, and several people whose job titles I have given up trying to understand. This is a wonderful problem to have, and a dangerous one. The first temptation , the one that hits churches in tech corridors hardest , is uncritical adoption. The reasoning goes like this: our people are in tech, tech is the future, therefore the church must be on the cutting edge of tech to remain relevant. So we build an app. We integrate a chatbot into our pastoral care system. We generate sermon outlines with GPT and quietly tell ourselves it's just for inspiration. The trouble is not that any of these tools is necessarily wrong. The trouble is that we have allowed the industry's tempo to set our own, and a church on someone else's tempo is no longer leading anything. The second temptation is the opposite, and it is just as bad: performative Luddism , the bishop who announces a ban on AI from the pulpit, the small group leader who refuses to use any digital tool on principle, the wholesale conflation of every new technology with the tower of Babel. This is theatre, not theology. It feels brave but it is mostly nostalgic, and it abandons the people in our pews who actually have to make decisions about these tools in their jobs on Monday. The serious theological response is neither. It is to form people who can use these tools without being formed by them, which requires the much harder work of forming people in the first place. That work is not new. It is the work the church has always done, badly and slowly, and it does not get easier just because the technology gets faster. ## What Slow Thinking Actually Looks Like on a Sunday So, practically. What does a congregation do? First, we keep reading the Bible together, out loud, in long stretches , whole chapters, whole books. The discipline of sitting under a text that does not optimise for engagement is now genuinely countercultural. Every algorithm in your pocket is designed to fragment your attention. The reading of Scripture is designed to reassemble it. If you do nothing else, do this. Second, we pray for and with the people whose lives are being reshaped by these technologies , the displaced workers, but also the engineers building the systems, who carry burdens we rarely acknowledge. I know engineers who lie awake at night wondering if their work is on balance good for the world. They do not need our hot takes. They need our prayer and our friendship. Third, we practise lament. The Psalms contain more lament than praise, which is statistically embarrassing to most contemporary worship sets. A church that cannot lament cannot face the actual griefs of an actual decade. Make space, regularly, for people to say out loud that something is not right and to bring it to God without resolution. Fourth, we cultivate friendships across class and capability. The economist and the drug dealer (to borrow an old image of mine), the engineer and the man who lost his data-entry job , these people belong at the same table, and almost nowhere else in the city will put them there. The church can. If we do not, no one will. Fifth, we teach, slowly and repeatedly, the doctrine of the image of God. Not as a slogan but as a daily lens: the person at the till, the person on the train, the person whose email you are about to send curtly, the person whose job is about to disappear. ## Eighteen Months, or Eighteen Years , The Gospel Doesn't Wait to Be Useful Here is the thing I keep coming back to. If Suleyman is right and human-level AI arrives in eighteen months, the church will still need to do what it has always done. If he is wrong and it takes eighteen years, the church will still need to do what it has always done. If it never arrives at all, in any meaningful sense, the church will still need to do what it has always done. Our calling is not contingent on the accuracy of his forecast. This is not because the forecast doesn't matter. It is because the things the church has to offer , lament, solidarity, resurrection hope, the slow formation of dignified human beings , are already needed. They were needed in Augustine's Rome and in Wesley's Bristol and in the East End during the Blitz. They are needed now, in a London where men in their fifties learn they are dispensable by email. I do not know what Mustafa Suleyman's eighteen months will bring. I doubt he does either, not really. What I know is that on Sunday morning, in a rented school hall, we will sing and we will read and we will break bread, and the man with the quiet face will be welcome at the table. That is not a strategy. It is the gospel, and it does not wait to be useful before it becomes true. "Be still, and know that I am God." Psalm 46:10. The instruction has not been updated.