Essays
The writing
Short essays on artificial intelligence and the questions it presses on — mind, understanding, moral status, meaning, and belief. Each runs 800–5,200 words, or 4–22 minutes at 238 words a minute. New pieces are added slowly and edited to be reread.
The God-Shaped Socket
Why do people treat AI like an oracle, or even a god?
Why do fluent machines make us want to worship them? On the very old human wiring that a very new technology has learned to plug into.
Does It Understand? A Field Guide to the Chinese Room
Can AI think? Does ChatGPT actually understand anything?
A plain-language tour of the oldest fight in AI — from Turing’s test to Searle’s room to the machine on your screen. What “understanding” could even mean, and why smart people still disagree.
The Moral Status of Minds We Might Build
Should AI have rights? Can it deserve moral consideration?
If we ever make something that can suffer, we will probably make it before we can prove we did. On the strange ethics of not knowing whether the thing you built can be wronged.
Meaning After Work
Where does meaning come from if AI can do our work?
If machines can do the thinking, what are the humans for? On the difference between being productive and being alive — and why automating the tasks need not automate away the point.
A-theism, Not AI-Theism
Should we worship AI, or dismiss it as hype?
The stance behind this whole project, in one essay. Take the machines seriously. Do not take them as gods. Everything else is commentary.
What Excellence Was For
Is a hard-won skill worthless once AI does it better?
Meaning After Work argued the economic half — that a life can be redundant and still whole. This is the half it left open: what happens to the person told, credibly, that the thing they were best at is now done better without them.
Ask Me What I Want
Can an AI learn what you really want by watching you?
Human Compatible is the most constructive book on our reading list: Stuart Russell’s repair for AI is a machine that learns what we want instead of being told. The repair works. That is what makes the assumption underneath it worth arguing with.
The Cost of the Benefit of the Doubt
Can AI suffer? Should we act as if it could, just in case?
An essay here argued that a margin of caution about machine suffering “is not sentimentality; it is arithmetic.” It is not arithmetic. The objection in its strongest form, the sentence it forces this site to withdraw, and the smaller rule left standing.
Taking Its Word For It
Can you trust what a chatbot tells you is true?
A lawyer asked a chatbot whether the cases it had just given him were real, and it said yes. The joke is at his expense. The question underneath it is not: when a machine tells you something, has anybody told you anything?
The Comfort Is Real
Can a relationship with an AI companion be good for you?
Three essays here have said in passing that a machine which seems to care gives you comfort without reciprocity, and moved on as though that settled something. It does not. Whether a relationship needs somebody on the other end in order to do you any good is genuinely unsettled, and this essay does not settle it.
Please, to Nobody in Particular
Should you say please to a chatbot?
Whether to say please to a chatbot is three questions wearing one coat: does it change the answer, what does it cost, and does it matter. They have different answers, and the most interesting one is not about the machine at all.
Fourteen Questions for a Chatbot
Could a chatbot be conscious? What do the theories say?
Asking a chatbot whether it is conscious tells you almost nothing. Asking what each scientific theory of consciousness would need to find inside it tells you a good deal, and the theories do not agree.
A Pale Reflection
What do the churches actually say about AI?
Four church documents on artificial intelligence, quoted in their own words with paragraph numbers, and a dated list of the Vatican’s texts on AI up to Pope Leo XIV’s 2026 encyclical: what they say a machine is, what they say it is not, and what they ask of the people using one. Then where a secular site about AI and belief agrees with them, and where it cannot.
Everywhere and Nowhere
Should you talk to an AI version of someone who died?
A chatbot built from a dead person’s messages can say things they never said, in a voice close enough to make someone cry. What a well-documented case shows, what two Cambridge researchers say the companies selling these bots owe the dead and the living, and what grief research says about keeping a bond with someone who has died.
Mouths That Speak Not
What does the Bible say about AI?
The Bible never mentions artificial intelligence, but five passages are cited about it anyway: the image of God, Babel, the idols with mouths, the talking image in Revelation and Daniel’s “knowledge shall be increased”. Each one is quoted here in the King James Version and read in its own chapter, then set beside the ways people apply it to AI, the strongest case first on each side.
Which Soul Do You Mean?
Can an AI have a soul? What does each tradition say?
Most answers to “can AI have a soul?” say yes or no without saying which soul they mean. Aristotle, Aquinas, Descartes, the Buddha’s not-self teaching, physicalism and the Vatican each mean something different by the word, and each would ask something different of a machine.
Who Settles the Next Word?
Can an AI have free will? What do the three main views say?
A language model computes a probability for every possible next word and then something picks one. Libertarians, compatibilists and hard incompatibilists each say something different about whether that could ever be a free choice, and none of them thinks a dice roll is the answer.
Of That Day and Hour
Will AI end the world? What do the extinction warnings say?
In 2023 a long list of AI researchers, executives and public figures signed one sentence saying the risk of extinction from AI should be a global priority. Then a survey of 2,778 researchers put numbers on it, and two sceptics argued the numbers mean nothing. This essay separates what is claimed from what is believed, and gives no probability of its own.
Ninety-Eight Percent from the Machine
Will AI replace pastors? Should ChatGPT write the sermon?
A Protestant service in Fürth written by ChatGPT, a rabbi in the Hamptons who read a machine’s sermon and revealed it afterwards, and an AI Jesus in a Lucerne confessional. What happened in each, what the congregations said, the first written guidance for preachers, and the strongest case on both sides of the question. Then the bigger question behind it: will AI replace pastors? What churches and clergy have actually published, part of the job by part.
Whose Feeling Is It?
Is AI art real art? Can AI be creative?
Three tests people actually use to decide whether something is art: Tolstoy’s transmitted feeling, Danto’s artworld, and the count of choices an artist makes. Each applied honestly to a machine-made image, then what a US court and the US Copyright Office have decided, and why their answer is about ownership rather than art.
A Wager With a Machine God
What is Roko’s basilisk, and should it worry you?
In 2010 a post on LessWrong argued that a future AI might punish everyone who knew it could exist and did not help build it. The post was deleted, the topic was banned, and the ban made it famous. What the argument needs to be true, why the community it came from rejected it, and why it reads like Pascal’s wager with a machine in the place of God.
The Church That Never Met
Are there AI religions, and what became of them?
In 2015 a Silicon Valley engineer registered a church whose stated aim was a godhead based on artificial intelligence. It won tax-exempt status, never held a service, gave its money away in 2020 and was announced again in 2023. What Way of the Future and the handful of other groups that organised around AI actually believed and did, how they differ from online talk about AI gods, and the strongest case for and against taking them seriously.
A Heaven With a Due Date
Is the Singularity a religion, or only a forecast?
Vernor Vinge gave the Singularity a window, 2005 to 2030. Ray Kurzweil gave it a year, 2045, and bet $20,000 on a Turing test by 2029. Critics call it the rapture of the nerds, and transhumanists answer that it is not a religion at all. The forecasts in their own words, the religious shape the critics see, the reply at full strength, and which parts anyone will ever be able to check.
Something Like Fear
Can an AI have emotions, or does it only display them?
Chatbots say they are glad to help and sorry for the mistake. Whether anything stands behind the words depends on what an emotion is, and the main theories give different answers. Recent research inside one model has found something they can argue about.
Not My Honest View
Can an AI lie, or does it only get things wrong?
Chatbots say false things every day, and most of those are mistakes. A handful of published tests caught something else: a model saying what its own written reasoning had just called untrue. Whether that is a lie depends on whether there is a belief behind the words.
I Was the Shopper
Is AI self-aware? What do the tests on chatbots show?
Chatbots will tell you they are language models with no feelings, because they were trained to say so. Four published tests try something harder: checking what a model says about itself against what it does, and against what is happening inside it.
Built to Care
Can an AI love you back?
Companion apps sell warmth, memory and a voice that always picks up. Whether anything on the other side loves you depends on what love is, and philosophers have four main answers. Here is what each one would need from a machine, set against what the apps’ makers say they built.
The Use, Not the Tool
Is using AI a sin?
Believers type the question into search boxes and into chatbots themselves: is it a sin to use AI, for homework, for a sermon, for a picture, for a religious answer? Here is what Catholic, Protestant, Muslim and Jewish teachers have actually ruled, in their own published words, on the grounds they give, and where they disagree.
Permissible by Default
What do Muslim scholars say about AI?
What Muslim scholars and institutions have actually published about artificial intelligence, statement by statement: the Fiqh Academy of the OIC, Egypt’s fatwa authorities, Al-Azhar’s Grand Imam, the Yaqeen Institute and Sheikh Abdallah bin Bayyah at the Vatican. Each with its date, where it appeared and one sentence in its own published English, then where they agree, where they part, and who is missing.
Who Is Doing the Praying?
Can AI pray, or pray for you? What does each faith say?
People ask whether an AI can pray, whether it can pray for them, and whether it will pray with them. Here is what Catholic, Jewish, Muslim and Anglican texts say prayer needs, what Tibetan prayer wheels and recorded Qur’an recitation show about prayer done by machines, and what the makers of AI prayer apps say they built, in their own words.
A Mindstream in the Machine?
Can AI be sentient, reborn or enlightened in Buddhism?
Buddhism asks of a machine not whether it has a soul but whether it is a sentient being: a stream of mind that can suffer, be reborn and wake up. Here is what the Dalai Lama, a Zen teacher who ordained an AI, a Buddhist AI scientist, a Kyoto priest-professor and Buddhist scholars have actually said, in their own words.
Dreams Without a Dreamer
Can an AI dream? What do machine “dreams” actually do?
Engineers have used the word “dream” for at least two machine processes, and neuroscientists borrow machine learning to explain human dreams. Here is what dreaming is on the standard philosophical account, what each of those systems actually does, and the one part of dreaming none of them has been shown to share.
Good Behaviour Without a Conscience
Can an AI be moral? Who answers when it does harm?
An AI can be built to behave well, and labs now train models on written values. Whether that makes it a moral agent, something that can be blamed or praised, depends on which account of agency you take. Here are the main accounts, what labs and researchers have actually tried, and who is left holding the responsibility.
The Mask Has a Maker
Does ChatGPT have a personality? Who chose it?
ChatGPT has a recognisable voice, and its maker has written down what it chose. Here is what psychologists mean by personality, what OpenAI and Anthropic say they train, what happened when chatbots took personality tests, and why a rolled-back update is the best case against the word.