For this sci-fi story/world hat I'm building, I'm introducing my androids which, while having intelligence closer to a human, are still just machines in nature. Theoretically one can destroy the body over and over again, they won't necessarily "die" unless it's fried or destroyed. Although it makes logical sense, death has to have some meaning one way or another, right?

They are also mainly powered through a unique battery system that stores energy from solar resources. It stores electricity for solar self-consumption, time of use load shifting, backup power, and off-the-grid use. So if they ever shut down, they can just regain that power as unlike humans, machinery don't usually die by losing energy.

So I've been wondering if there would be anyway to have my androids die without both of their features having a easy fatal flaw that can be exploited?

EDIT: Sorry if I didn't make myself clear, but when I mean "killed," I meant in the "physical" sense; gunfire, explosions, etc. How, or more importantly, would they die if they died out of malice or misadventure?

  • $\begingroup$ Are you aiming for natural causes here? ie. excluding malice or misadventure $\endgroup$ – Pingcode Jun 20 '19 at 10:30
  • $\begingroup$ What is an android? Is the identity of the android linked to its physical body or is it linked to its software and data? This is important. Unlike humans, computers really do have immortal souls, which can be backed up and restored. With a little care and application of best practices, computer programs and data are as close to immortality as any divinity. $\endgroup$ – AlexP Jun 20 '19 at 11:13
  • $\begingroup$ This is the Ship of Theseus question applied to humans (and, by extension, androids). So, what in your world makes an android that specific android? It would behoove you to read (not watch) Asimov's robot short stories, specifically the Bicentennial Man. $\endgroup$ – RonJohn Jun 20 '19 at 21:29

Memory Deletion/Corruption

You may be aware of the horrifying fridge logic behind teleportation. Essentially, you get murdered and your clone walks out the other side.

But an Android's mind is not stored in a brain, but on a hard drive. You can move this drive to a new body if the original is too damaged, but if you were to make a copy of the Android's code and delete the original, you have just done the exact same thing as destructive teleportation.

Thus, any damage to an Android's hard drive, or more specifically the code on it, is permanent. They may have the ability to expand to fill available space, but anything that gets deleted is gone for ever.

Therefore, natural deterioration of an Android's hard drive, or damage to it due to accident or malicious act could in theory kill your Android.

If you want to run an experiment on this, take a program on your computer that you don't want any more, delete a random subset of files from it and try to run it. At best, it'll sort of work but complain about missing files. At worst, it is now inoperable.

A more exciting one would be to drill a couple of holes in your hard drive then try and use it. I do not recommend this with anything you want to use ever again.

  • 3
    $\begingroup$ Remember also that most Hard Drives and RAM sticks have a limit on how many times any individual bit can be rewritten before it starts to "stick" and become corrupt - think of it like writing on a sheet of paper in pencil and rubbing it out. Eventually, the paper wears thin and can't be reused.... $\endgroup$ – Chronocidal Jun 20 '19 at 11:11
  • $\begingroup$ It might be also worth noting that current HDD/SSD lifespan is much shorter than average human lifespan. Simply making the drive impossible to reach without damaging it (for example the drive embedded in some material making it water/weatherproof with different layers on top of that for physical damage protection - durability reasons during their lifespan - try to pull that apart and you will most likely end up with a damaged drive). You can enchance your world with very shady specialists that are able to extract the drive, alas with only a few percent chance of success. $\endgroup$ – Rachey Jun 21 '19 at 8:30

Hardware or software failure of the data storage system, until the operating system can no longer successfully reboot. That could mean the processor, disks, ram, etc.. is damaged from radiation, heat, explosions... whatever. Computers are sensitive. They need thermal regulation, EM shielding, shock absorption, lots of things to survive. Kind of like the brain.

Supposing the entire computer was physically destroyed (wherever they keep that, their heads?) and there was a data backup then that could be flashed into a new body and it would wake back up remembering only what happened up to that backup. But that backup file could be vulerable to data corruption or physical destruction too.

Death in a final sense would mean the main computer and all backups are irretrievably destroyed.


Volatile Memory. When your computer is running, it loads data into RAM - this is volatile memory, and only maintains data for so long as it is powered. It also tends to be much faster than using the Hard Drive - non-volatile memory, which can maintain data without power.

(I should point out here that a Solid State non-volatile drive is far faster than a mechanical Hard Drive, but still not quite as fast as RAM)

For this reason, many industries often run software directly in RAM without the computer's Hard Drive. A good example is Banking - anti-fraud software often runs in RAM because it needs to analyse your card transaction and decide whether to flag it as fraud (and block the transaction) before the authorisation is sent back to the till!

So, some underlying aspect of your Android's programming is required to run in RAM. For special circumstances (maintenance, et cetera) this can be temporarily backed up (like putting your computer into "Hibernate") - but an unexpected loss of power would wipe it.

You might be able to repair the android, recharge it, and turn it on again - but it would be back to factory settings. No personality, no memories, just a newborn with the face of your friend...

  • $\begingroup$ There have been computer architecture with single level storage since forever. The currently commonplace distinction between data in memory and data on disk is an historical accident due to the dominance of operating systems (such as Unix, VMS and Windows) which make this distinction. For an example of an operating system which does not distinguish between volatile and persistent storage consider the venerable OS/400 and its descendats. (The trick is to operate transactionally.) Long story short, having volatile data is a choice and not a necessity. $\endgroup$ – AlexP Jun 20 '19 at 11:20
  • $\begingroup$ Daily backups to solid state. Power outages just reset the android to its state this morning. No self respecting engineer would do any less. $\endgroup$ – Muuski Jun 20 '19 at 20:52

Your android uses a self pruning/growing neural network implemented inside of a quantum computer in order to control its actions. Its state cannot be easily copied as it exists as state possibilities, inside of a quantum net.

The problem with using a quantum computers is that eventually the qbits that make them eventually destabilize and they become decoherent (death). Modern Quantum computers are kept at ~0K so that they last as long as possible, but they do not have a long lifetime. You would have to hand-wave some material or other method that allows you to create a long lasting room temperature stable quantum computer.

  • 1
    $\begingroup$ This is the only answer that addresses backups, which is, in my opinion, what the question is really about. $\endgroup$ – Muuski Jun 20 '19 at 21:01
  • $\begingroup$ Thank you. I also think that having it run using a probability calculator rather than a deterministic Von Neuman machine would explain why each of the androids could be considered a unique individual rather than a clone. $\endgroup$ – Garret Gang Jul 10 '19 at 20:26

Heat death of the universe

Though if they're bound to a single solar system, then a nova would do, or anything planet destroying if they're bound to a single planet.

It's very much a game of remote redundant backups, you can only destroy an android if you destroy the android and all backups.

  • If you give them a highly redundant system then they're pretty much impossible to destroy.
  • If you put all their backups in one building with no external network connections that they have to return to to recharge, then you only have to destroy that one building with them in it.

If a human's brain gets destroyed by a bullet passing through, and then we transplant a new brain into the same body, then we would still consider the original owner of the body dead. The new brain has different memories, experiences and capabilities.

If an android's memory and computational equipment is destroyed then we should also consider it dead, unless he's got an up-to-date backup it can use to put on the new chips. Even so the android can still assume that he/she/it is dead, and that a clone with his memories inhabits his old body if that happens. So to the android death would mean the same as it is for us: if the things that make you YOU are destroyed, you are dead.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.