What is the simulation hypothesis?

Formalized by Oxford philosopher Nick Bostrom in 2003, the simulation hypothesis is the argument that it is statistically probable our reality is a sophisticated computer simulation run by a more advanced civilization. The strength of the argument is not that it proves we are simulated — it's that nothing in our experience could falsify the possibility.

What is a controlled hallucination?

A term popularized by the neuroscientist Anil Seth: the brain has no direct access to the outside world, only to electrical signals from the senses. It generates a best-guess model of reality from those signals, and we experience that model. The 'controlled' part is the brain's constant updating of the model against incoming sensory data.

What can't be simulated?

_morfema's thesis: emotions, in the moment of having them. You can be deceived about what caused a feeling, what to do with it, even whether what triggered it was real or synthetic. But you cannot be deceived about having had the feeling itself. The act of feeling is the one part of experience that resists simulation from the inside.

note · v

the simulation and what resists it

the line between what's real and what isn't has not just blurred. from this side of it, it has already been crossed. the media you are looking at right now might be ai-generated. the messages you reply to might be too. the photos, the voices, the songs, the conversations — any of it could be synthetic. and there is no infallible test you can run from inside your own perception to distinguish the synthetic from the unmediated.

nick bostrom argues it is mathematically probable that the entire substrate of your experience is a simulation run by a post-human civilization. the point is not whether he is right. the point is that nothing in your daily experience could prove him wrong. the image on your retina is reconstructed. the world outside your skull is, in the neuroscientist anil seth's phrase, a controlled hallucination. the self is a model your brain is building of itself in real time, and updating constantly to keep the model coherent.

once you accept that everything visible can be simulated, only one question is left: what cannot? this project is built around a single answer. emotions are all that's left as real.

not because emotions are metaphysically special. because they are the only thing in this whole picture that you cannot be wrong about while you are having them. you can be wrong about what triggered the feeling. you can be wrong about what to do with it. you can be wrong about whether the song that made you cry was made by a person or by software. but you cannot be wrong about having cried.

the whole record is the hypothesis. your response is the data.

once everything visible can be simulated, only one question is left: what cannot?

← all notes