Mind uploading almost always gets framed as a story about living forever. Scan a brain, run it on a computer, and the person who scanned it never has to die. That framing is intuitive, and it's also the wrong question. The actual conceptual crux isn't survival. It's what happens once the same mind can exist as more than one thing at the same time.
If a scan produces a working copy while the original brain is still alive and thinking, there are now two things with an equal claim to being you, diverging from the same starting point in different directions from the moment the copy is made. Immortality was never really the hard part. Multiplicity is.
How close any of this actually is depends on who you ask, and the honest answer is: much further than the framing suggests.
Having the map has never been the same thing as running the mind.
How close are we, actually?
Scientists have had the complete wiring diagram of a roundworm's brain since 1986. Just 302 neurons. Nearly 40 years later, nobody has successfully run a working simulation of it.
Progress on more complex brains is real, just nowhere close to human scale. Researchers have completed the full wiring diagram of an adult fruit fly's brain, roughly 140,000 neurons, and a separate project mapped a fly's brain together with its full nerve cord, about 166,000 neurons, the first wiring diagram spanning a brain and its spinal-cord equivalent. For mammals, the state of the art is a single cubic millimeter of mouse cortex and a single cubic millimeter of human cortex, roughly 57,000 cells and 150 million synapses. A human brain has about 86 billion neurons.
- C. elegans (roundworm) — 302 neurons, fully mapped since 1986, still unrun
- Adult fruit fly — ~140,000 neurons, fully mapped
- Fly brain + nerve cord — ~166,000 neurons, fully mapped
- Human cortex sample — 1 cubic millimeter, ~57,000 cells, ~150 million synapses
- Full human brain — ~86 billion neurons, not yet attempted
One separate project, MICrONS, has mapped 200,000 cells and 523 million synapses combined with real-time activity recordings from more than 70,000 neurons, genuinely impressive, still a rounding error against a whole brain. Costs are falling fast: reconstructing a single neuron cost roughly 16,500 dollars in the original C. elegans project and closer to 100 dollars in recent zebrafish work. Even at that improved rate, mapping 86 billion human neurons remains a project of an entirely different order of magnitude.
The most thorough assessment of the field, the State of Brain Emulation Report, published in October 2025, is the first comprehensive reassessment since the original 2008 roadmap. Its conclusion: human whole-brain emulation is at minimum 30 to 40 years away. Many neuroscientists working in the field privately estimate longer, or question whether it's achievable at all. Public forecasting communities land in a similar range, with median guesses stretching from the 2060s to the 2080s, and a real share of forecasters assigning meaningful probability to never.
A Harvard-MIT researcher working on the problem, Davy Deng, has described a striking geographic split in how seriously people take the idea at all: mention whole-brain emulation in Silicon Valley, he's said, and people lean forward. Mention it in Boston, among working neuroscientists, and the reaction tends to be closer to suspicion.
Why is "who wakes up" the harder problem?
Even solving the scanning problem completely wouldn't settle the actual question. A structural copy raises an issue no better microscope can resolve: whether that copy is a continuation of the original person, or a new person who simply believes they are that person, carrying the same memories with none of the same claim to being the one who had them.
This isn't a purely hypothetical debate anymore. A company called Nectome has published research on preserving a whole brain at synaptic resolution using a procedure compatible with physician-assisted death, which is legal in Oregon. Whatever the eventual technical timeline, the policy question of who gets to authorize a brain's preservation, and under what circumstances, is already live.
What does "copyable" actually change?
If the timeline runs anywhere close to the 30-to-40-year floor the 2025 report describes, the civilizational question is not really about individual survival at all. It's what happens to a species once being one person, singular, is no longer something a brain guarantees by default. Institutions built around a person having one body, one legal identity, one continuous stream of memory and liability, would face a genuinely new category of problem the moment that assumption stopped holding, not because anyone chose it, but because the technology made it possible.
Almost every civilizational structure currently assumes a person is singular by default. A vote, a contract, a marriage, a prison sentence, an inheritance, all of it presumes exactly one continuous claimant. None of those systems were built with branching in mind, because until very recently branching was not a physical possibility worth planning around. Whether that assumption holds is a genuinely open question, not a settled one, and this publication has no answer to offer beyond naming that the question exists and is worth taking seriously before the technology forces it.
None of this is close enough to demand an answer yet. The gap between a worm's brain, mapped for 40 years and still unrunnable, and a human brain, mapped in single cubic millimeters at a time, is the honest measure of how far away that day actually is. But the questions underneath it, whether a copy is you, and what a civilization does once one person can become more than one, don't require the technology to exist yet. They only require someone to have asked what happens if it eventually does.