Post 1, intro/general

Post 2, flight enabled by plasma thrusters

I had mentioned the 'Singularity' in a previous version, but that was incorrect. Fully automated luxury communism is much more accurate.

To make personality/identity/memories survive physical destruction, we need to scan the exact state of a brain. Then a replacement can be 3-D printed from cloned cells and electrically/optogenetically stimulated until it has your connectomic pattern. Now if you get eaten by a Tyrannosaurus, you carry on from your last save-point.

A word on philosophy is necessary here. Some philosophical theories would say "that's just a copy, not a continuation! It's a different person, albeit an identical different person." For the purposes of this post, let's say Zhuangzi's philosophy is correct; the feeling of I-ness is the same everywhere it occurs. It's modified by different specifics, different memories and cognitive tendencies, but the sense of 'I' is universal. It therefore follows that an exact copy of a brain will be the same person; there is no unique identity that could be lost in copying; the thoughts, memories, tendencies, and the universal sense of 'I' are all there is.

First we need a perfect scan of your brain; that's the hard part. It's dark in there, and electrical signals all overlap and get lost in the noise.

Destructive brain-scanning is by far the most realistic

There's actually one very plausible way to scan the exact state of a brain. We cut the brain into very thin slices, and scan each with an electron microscope.

Those of you who know a lot of science might have spotted an issue. Slicing someone's brain into small wafers to feed into a scanner can have adverse effects on that person.

You would be dead for a while while your brain is analysed, and a perfect copy is made. At the end you walk out merrily, knowing your backup is now saved on the cloud with redundant backups. Current techniques would take weeks to scan a brain, but this will surely speed up in the future.

This would be one option for fictional immortality. The issue is: dicing up brains is extremely gothic. It was supposed to be a hedonistic utopia, not the surgical-horror genre. Vibes and science both matter for worldbuilding. I could go for the artistic fix: this backup system gives citizens a confident attitude towards Kali herself. That's the image at top: he is coming out of the clinic laughing about his death because he's beaten it. You celebrate your backup date, go in to the procedure, and come out happy your soul is safe. I could say it takes three days. The idea of a society where every joe routinely dies for three days then comes back to life having conquered death is religiously very funny.

Non-destructive perfect brain-scanning will require a bunch of microendoscopes

You may think you could scan a living brain from the outside, with a helmet. Put that out of your head. It can't be done, not to the level of detail we need here.

It must be done invasively. The challenge is to make tiny probes (microendoscopes) that run throughout the whole brain and scan from the inside. This is exactly the same idea as the neural lace from books I haven't read:

• The scanner must have spatial resolution good enough to map each neuron. (Perhaps even finer spatial resolution is required, scanning dendritic spine structures; the truth is we don't know how the brain works, so we don't know what information is relevant.)
• The scanner must have the temporal resolution to track each firing-event, in other words 1-2 millisecond framerate.

We can scan the brain thoroughly by putting in loads of things, but then 5% of your brain is gone and you're in trouble: (blue area represents scanning range)

So make the probes really small? But now we have unscanned brain regions, and the point was to get a complete map of the brain:

We must balance two conflicting goals: get probes 'close enough' to every part of the brain, while leaving as much brain tissue as possible alone/minimising the total volume of the neural lace. That link from before about neural lace says it occupies less than 0.5% of the brain volume.

Optical scanning can get to 1.6mm

The first decision to make is what scanning modality: electric or optical. The electric ones like Neuropixels allow very miniaturized shanks, which is good but ultimately they just can't scan far enough away from the tooltip. Electrical waves caused by neuronal activity get lost in background noise within 0.1mm. If you tried to scan the brain electrically, you end up with this again –

Photons travel further, so the neural lace will have to be a bunch of optical microendoscopes. These also have the advantage of seeing dendritic structures, which electrodes do not.

The paper 'In vivo two-photon microscopy to 1.6-mm depth in mouse cortex' shows photonic microscopy can go deep. 1.6mm is an impressive penetration depth (at least that's what I tell my girlfriend).

We can double that

Two opportunities for massive improvement...

• A paper called 'In Vivo Deep-Brain Structural and Hemodynamic Multiphoton Microscopy Enabled by Quantum Dots' reported 2.1mm deep scanning by replacing the photonic flashes with quantum dots. (Quantum dots are nanoscale crystals (they're not quantum-sized; misleading name) that trap and emit photons in special tunable ways.)
• Most challenges become much easier if you take a moment to clear your head first. 'In Vivo Optical Clearing of Mammalian Brain' on biorxiv shows how some simple non-toxic molecules can make the brain more transparent to light temporarily. And their gains are huge: they report the signal-to-noise ratio improved by 385%. That means 2.2× scanning distance, from 2.1mm to 4.6mm (signal falls off with distance (inverse-square) so distance scales to the square root of signal-to-noise ratio.)

How big is the probe?

A paper from 2011 with DOI 10.1117/12.874883 describes a side-viewing, all-optical probe 250 micrometers in diameter.

I would have preferred much smaller; human hair is 50 to 100 micrometers. However this is a hard limit; if you miniaturized any more there wouldn't be enough photons to count. Electric scanning would have allowed something much thinner, but has a worse scanning depth (and ultimately a worse ratio of probe diameter to scanning depth).

How big is the neural lace?

We said we could get scanning depth of 4.6mm. In theory you could have a space between probes 2× their scanning depth (9.2mm); their ranges would meet in the middle. In practice, you should allow a good overlap. Let's put them 6.8mm apart.

If the distance between probes is s, and probes are packed hexagonally (the correct choice of geometry when covering a field with radii), then each probe is responsible for scanning an area ≈ (√3/2)·s²

• If the probes are 6.8mm apart, one probe per 40mm² That's the area of unpoked tissue.
• And the area of the pokers: the diameter (as we already saw) was 0.25mm, so r 0.125mm, and πr² = 0.05mm²

So the lace occupies about 0.125% of the brain.

We said each probe scans a cross-sectional area of 40mm². The human brain is less than 320× that in cross-section, so picture about 320 threads, each 2½-5 times thicker than a human hair (depending on what kind of hair you have lol), running through your brain back to front, and meeting at an external plug behind your ear that you'll use for backups.

Snapshots, not life-logging

Don't mistake this for something that reads all your thoughts all the time. It enables you to go in for a backup. Maybe you go in for a backup every year on your birthday. Your memories are backed up to that point and if you are thrown into a vat of acid, you'll (subjectively) revive at the backup clinic.

Reasons it takes a few hours once a year –

• The photons have to be excited by shooting infra-red lasers into the brain. This will inevitably generate some heat. The backup-technicians do it bit-by-bit to avoid cooking the brain. Your head has water running over it throughout to remove heat. This is part of the strange but strangely pleasant experience, like getting your hair washed professionally at the salon.
• You have to get injected with the quantum dots
• You have to get take the brain-clearing chemicals.
• Light inside the brain has chemical effects that are themselves toxic, so a few hours exposure a year is best.
• Brain activity has to be logged for a few hours to give sufficient data. When we are measuring single neurons firing from 'distant' (3mm or so) probes, we are counting photons on our fingers. We need a decent logging window to make valid inferences with good p values.
• Some neurons are dormant, won't naturally fire during a few-hour logging window. So these get stimulated by external non-invasive brain-stimulation (doesn't have to be neuron-level precise). This adds to a trippy experience.

This is great news for privacy; it's not a thought-tracking technology, it's a backup-storage technology.