Minds, copies and machines · part 3 of 4 · by Bill Guan
If you can be copied, the copy will be just as sure it's you
The big question
If a machine made a perfect copy of you, which one would be you?
The puzzle is older than computers, and it gets less hypothetical every year. You're about to meet Parfit's teleporter, the thought experiment that still splits philosophers, and see how close brain mapping has come to turning it into a question for engineers.
Here's a thought experiment the philosopher Derek Parfit made famous. You step into a teleporter. It scans you down to the last molecule, destroys your body, and sends the blueprint to Mars, where a machine builds an exact replica 1. The person who steps out on Mars remembers your childhood, loves your family and finishes your sentences.
Did you just travel to Mars, or did you die while someone new woke up? Now make it harder. Suppose the machine malfunctions and leaves the original intact, so there are two of you, one on Earth and one on Mars, each completely certain they're the real one.
Your guess first
A fruit fly's entire brain has about 50 million connections between its neurons. How many did scientists find in about one cubic millimeter of mouse brain?
That speck of mouse brain held 523 million connections, about ten times the whole fly brain, and mapping it filled 1.6 petabytes of data. Nobody can copy a brain yet, but the scanning half of the job is moving fast.
Copying a brain is turning into an engineering question
Until recently, copying a mind sounded like pure science fiction. The recipe for what's called whole brain emulation goes roughly like this: preserve a brain, slice it, scan it, rebuild it in 3D, then simulate it. Anders Sandberg and Nick Bostrom's 2008 roadmap describes the goal plainly: "take a particular brain, scan its structure in detail, and construct a software model of it that is so faithful to the original that, when run on appropriate hardware, it will behave in essentially the same way as the original brain" 2.
The scanning part has moved fast since then. In 2024 scientists published the first complete wiring diagram of an adult fruit fly brain, 139,255 neurons joined by about 50 million connections 3. In 2025 a team mapped about a cubic millimeter of mouse brain, with more than 200,000 cells and 523 million connections, taking up 1.6 petabytes of data 4. And in 2026 a startup reported a simulated fruit fly brain driving a virtual body with "91% behavior accuracy," a company claim that still awaits peer review, from a model that can't form new memories 5.
A whole mouse brain hasn't been mapped yet, and one research group projects an optical method could do it in about five years 6. A human brain is vastly bigger again, but the direction is clear, and the question Parfit asked as a puzzle is slowly becoming one engineers will face.
Two copies, each sure it's the real you
Philosophers have argued about this for a long time, and there are roughly two camps. One camp says you are your body, or at least your brain, so the teleporter kills you and the replica is a new person with your memories. The other says you are the pattern, meaning your memories, personality and habits of mind, so as long as the pattern continues, you continue.
Parfit's own answer was more radical. He argued that what matters in survival is psychological continuity, the unbroken chain of memories and personality, and that the question "is it really me?" carries less weight than we feel it does 1. The Stanford Encyclopedia of Philosophy sums up his view in a single line: "Identity itself has no practical importance" 7.
Replace the brain one neuron at a time
There's a gentler version that I find harder to dismiss. Instead of copying your brain all at once, replace it one neuron at a time. David Chalmers imagined exactly this: "only a single neuron is replaced. Its replacement is a silicon chip that performs precisely the same local function as the neuron." Then another, and another, until "every neuron in the system has been replaced by a chip" 8.
Chalmers used this to argue that your experience would most likely carry on through the whole process, though he admits his reasoning doesn't "constitute conclusive proof" 8. It leaves open whether the person at the end is still you, yet it makes the gradual path feel very different from the teleporter, even though the end result is the same.
Digital minds could trade memories and switch each other off
Digital minds could do things no human can. They could share experiences directly, download each other's best memories and delete the dull ones, with some kind of barrier between minds so that each keeps a sense of being an individual. And there's a darker question for anyone who merges with machines: could someone else delete your copy, or simply switch you off? That last question is where copying minds stops being philosophy.
Whoever runs the hardware controls the copies
My assessment: the moment minds can be copied, whoever controls the hardware controls the minds running on it. A copy can be paused, edited, duplicated a thousand times for a job, or deleted when the job is done. If those copies are conscious, we'd face what Nick Bostrom calls mind crime (simulating minds that can suffer and deleting them once they're no longer useful) on an enormous scale, and if they're copies of real people, consent becomes a nightmare of its own. Do you get a say over what's done to your copy? Does your copy?
We don't need to settle Parfit's puzzle to see the practical rule that falls out of it. A copy will feel exactly as real to itself as you do to yourself, so any world that makes copying cheap needs rules, written in advance, about who owns a mind and what can be done to it, and the best time to write them is before the first scan of a human brain, while we can still argue about it as a thought experiment.
A mind that suspects it might be inside a simulation could act as if it's always being watched, and today's AI models already seem to notice when they're being tested.
Sources
- Carsten Korfmacher, "Personal Identity," Internet Encyclopedia of Philosophy (summarizing Derek Parfit, Reasons and Persons, 1984). iep.utm.edu/person-i/
- Anders Sandberg and Nick Bostrom, Whole Brain Emulation: A Roadmap, Future of Humanity Institute Technical Report 2008-3. ora.ox.ac.uk/objects/uuid:a6880196-34c7-47a0-80f1-74d32ab987…
- Sven Dorkenwald et al., "Neuronal wiring diagram of an adult brain," Nature (2024-10-02). pmc.ncbi.nlm.nih.gov/articles/PMC11446842/
- Allen Institute, on the MICrONS Consortium's Nature papers (2025-04-09). alleninstitute.org/news/scientists-complete-largest-wiring-d…
- Michael Andregg, Eon Systems, "We've uploaded a fruit fly" (2026-03-08). eon.systems/updates/weve-uploaded-a-fruit-fly
- E11 Bio, "Roadmap" (2024-12-03). www.e11.bio/blog/roadmap
- "Personal Identity," Stanford Encyclopedia of Philosophy (rev. 2023). plato.stanford.edu/entries/identity-personal/
- David Chalmers, "Absent Qualia, Fading Qualia, Dancing Qualia" (1995). consc.net/papers/qualia.html