In the near future, humans will have manufactured machine sentience. Such a consciousness, the first of a new species, must operate on Earth.


A memristor-based human brain analogue, which seems the most likely candidate for sentience, requires about 700W peak power to operate and requires the area of a medium-sized pizza. This implies having sufficient batteries to supply the brain with continuous electricity. To date, double carbon batteries seem to be a marked improvement over any type of lithium-based battery.

I haven't performed the calculations, but it seems reasonable to assume that such a species would not use flight to locomote, due to the weight of its power supply.

(Solar powered flying machines exist, but they are huge, so they wouldn't fit inside a house, making it difficult to socialise with humans. Detachable [or foldable] wings and re-purposed electric motors for ground travel is possible, but that cascades into other problems that probably make it an impractical design.)

First Draft

Obvious basic requirements seem to be:

  • Cameras (360 degree perspective, infrared, ultraviolet)
  • Speaker (broad frequency range)
  • Microphones (highly sensitive)
  • Arms (multiple)
  • Hands (because our objects are built for human-like hands)
  • Track-legs (locomotion, staircases, and height change)
  • Internal sensors (temperature, barometer, gyros, compass)


What appendages or abilities would such a species need, if any, to: maneuver in our world, manipulate physical objects, build children (hence evolve), and interact with humans?

In other words, if you could design an optimal inorganic sentient species without being held back by Earth's evolutionary baggage, what would it need?

Here is a render of the first draft requirements (hands aren't drawn yet and a few other items are missing, but it serves to communicate the idea):

Sentient Species v1

Additional Resources

  • $\begingroup$ Does it get to rely on the fully fledged infrastructure of a society, or does it have to be able to repair itself and replicate just with what it has (like organic life does)? $\endgroup$ – Cort Ammon Nov 3 '15 at 22:30
  • $\begingroup$ It can leverage our resources and infrastructure. That is, it can talk with humans to request parts. (Doesn't mean humans have to obey, but the possibility is certainly there.) Eventually, since it is sentient, it will want the ability to repair itself (i.e., less reliance on humans to fix it). It'll still need objects we've manufactured, however. $\endgroup$ – Dave Jarvis Nov 3 '15 at 22:44
  • 1
    $\begingroup$ I'm not 100% sure that 'evolve' is the word you're looking for here. If the machine builds it's offspring that suggests it has an active role in the design of the next generation, which isn't evolution (it's 'genetically' modified designer babies). If that's the case then it doesn't really matter what you give this generation of device, two or three generations down the line you'll have a device with the ideal set of requirements. $\endgroup$ – Joe Bloggs Nov 4 '15 at 11:59
  • 1
    $\begingroup$ If you want your sentient machine to operate in a world optimized for human use, it should be at least moderately humanoid. that means, its cameras should be somewhere near human eye level, its height should be somewhere nea a humans height. Its arms and hands should be able to operate in regions where human hands can. Thus they would be able to use a car, reach the upper shelves in buildings, navigate stairs and doors, and generally do everything a human can do. $\endgroup$ – Burki Nov 4 '15 at 14:11

Use a remote body.

It's not "being held back by Earth's evolutionary baggage" like, for instance, needing to carry its brain along with its body. The obvious answer for a creature with a massive brain requiring a significant amount of power to survive is to keep that brain near a reliable power source. This is especially obvious if that creature has natural access to wireless technology.

This is beneficial in several ways:

  • The creature is not killed when its body is destroyed
  • The brain can be housed near a reliable high-power source, like a hydroelectric dam or nuclear power plant.
  • The creature can have multiple modes of motility for different situations (legs, wheels, wings, and impellers)
  • It can effectively travel instantly from one place to another when moving from one body to another.

All sensor data is already coming into the brain digitally, making that link wireless is not a difficult step. If I had my brain transformed into a computer, I would not be counting on battery technology to keep me alive, and given the option I don't think these creatures would either.

  • $\begingroup$ Note that a memristor only requires power to change state (i.e., think). No power doesn't mean death. Not sure if a medium-sized pizza qualifies as "massive," nor 700W (seven tungsten light bulbs) as significant. The idea of a remote body is quite good. It adds some interesting ideas to explore (i.e., what does it mean to move from body to body when it can control multiple bodies simultaneously; and what happens when the wireless connection fails). Very cool, thank you. $\endgroup$ – Dave Jarvis Nov 5 '15 at 0:52
  • $\begingroup$ @DaveJarvis I'm familiar with memristors. The massive aspect refers to the specific energy of existing battery technology. Lithium-ion batteries have a specific energy of 1.8 MJ/kg. Ignoring losses due to conversions or high current draw (which would be significant), at 700W a one kilogram battery will last 0.7 hours. With those losses, and powering the body to move around, you're looking at a significant portion of the creature to be batteries. If a body runs out of juice, no big deal, but if it has the brain in it then the life is at the mercy of the environment/society. $\endgroup$ – Samuel Nov 5 '15 at 0:54

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.