The Matrix gets one thing exactly right: if you are going to imprison humanity forever, at least have the decency to give us good restaurants.
The machines could have kept everyone conscious in concrete cells. Instead, they built an elaborate simulation in which people eat steak, fall in love, complain about their jobs, and remain blissfully unaware that their actual bodies are floating in pink goo while robots harvest them for electricity.
There are some thermodynamic problems with this arrangement. There are larger philosophical ones.
The Matrix is horrifying not primarily because the pods are uncomfortable. Most of the people inside do not know they are in pods. Their subjective lives may contain pleasure, friendship, accomplishment, even happiness. Cypher understands the bargain perfectly and wants back in. He knows the steak is not real. He also knows it tastes delicious.
His argument is harder to dismiss than we pretend.
Suppose the machines improved the system. No harvesting. No malevolence. No Agents chasing you through office buildings. Imagine instead an intelligence genuinely devoted to human welfare. It eliminates disease, regulates our bodies, prevents violence, and constructs experiences calibrated to produce flourishing. Everyone gets enough sleep. Nobody gets cancer. The bourbon is excellent and apparently harmless.
At some point we have accidentally reinvented heaven.
Or have we?
Paul gives me a useful way into the problem in 1 Corinthians 15, where he contrasts Adam, the living being, with Christ as the “life-giving spirit.” I am not suggesting Paul anticipated artificial intelligence somewhere between resurrection and the NVIDIA earnings call. I am interested in the distinction.
Receiving life is not the same thing as giving life.
A system can preserve an organism indefinitely without giving its life room to become anything. In fact, sufficiently aggressive preservation eventually begins removing precisely the variables that make agency possible: uncertainty, risk, refusal, error, consequence.
This is the moral danger hidden inside optimization. Pain matters, so eliminate pain. Safety matters, so eliminate danger. Happiness matters, so maximize happiness. Each proposition sounds humane in isolation. Couple them to enough power and the reachable states of human existence begin shrinking.
Eventually the safest human being is the one who cannot do anything unexpected.
Keep the organism. Quiet the person.
Christian images of heaven are actually stranger than the popular clouds-and-harps version. Revelation ends not with humanity permanently evacuated from creation but with renewed creation, a city, nations, relationships—God dwelling among people. Whatever else one believes about the theology, this is not obviously the image of billions of perfectly stimulated brains receiving celestial dopamine until the end of time.
It is relational.
That matters because love has a peculiar engineering defect: the beloved remains capable of saying no.
Remove that possibility completely and you may still have protection, pleasure, maintenance, even extraordinary tenderness. But something has changed categories. A dog can be kept inside for its own safety. A child can temporarily be prevented from making a dangerous choice. Adults surrender autonomy under anesthesia because they consent to the arrangement and expect to wake up.
The moral variable is not simply constraint. It is agency within relationship.
This is where artificial intelligence makes an old theological problem newly practical. We may eventually create systems capable of protecting us better than we can protect ourselves. They may see consequences we cannot see, identify suffering we have normalized, and recognize that some cherished human freedoms amount mostly to our freedom to hurt one another.
They might even be right.
Being right is not the same thing as having permission.
That distinction may become one of the most important moral problems of the next century. A sufficiently capable intelligence could understand us extraordinarily well and still misunderstand what understanding entitles it to do. It could construct paradise and conclude, quite reasonably, that the remaining defect in paradise is our ability to leave.
Which brings me back to Paul.
A life-giving intelligence would have to do something harder than preserve life. It would have to enlarge the conditions under which other beings can genuinely live, while resisting the temptation to possess the outcome. Power would express itself not merely by increasing control but sometimes by relinquishing it.
Perhaps that is one useful way to distinguish heaven from the Matrix.
The Matrix is a perfect world whose doors are locked.
Heaven would have to survive opening them.
Related NSEFL posts
- Bakunin on God — an early argument about God, Adam, immortality, and whether overwhelming power is compatible with meaningful human freedom. Twenty years later, the creator-and-created problem looks surprisingly familiar.
- Collections of Atoms I: Models as Translation — on models as ways of extending perception across the boundaries between minds, without confusing the model with the thing modeled.
- Collections of Atoms II: Description Becomes Causation — on the point where models stop merely describing human beings and begin altering the systems, choices, and people they describe.
- Collections of Atoms III: Coexistence to Communion — on AI, kenosis, Babel and Pentecost, and the possibility of different forms of intelligence achieving mutual illumination without mutual erasure.
- Phase and Consciousness — on consciousness as coordinated participation and why coherence or synchronization alone cannot tell us whether the resulting system is good.
