LogFAQs > #903253823

bulbinking posted...

7, 4, 3, 1 are also impossible.

1. My computer automatically increases the speed of its CPU and GPU fans in an effort to maintain a constant internal temperature. This one's easy even without looking at hypotheticals; many automatic systems currently in use have the ability to regulate themselves to maintain optimal operating conditions.

3. Metabolism is the process of taking in material and converting it to energy. Most robots do not do this internally, instead relying on external power sources, but a robot that relies on solar power is effectively metabolising (it's functionally analogous to photosynthesis, it just skips a few steps). Hypothetically, a steam-powered robot that burned coal to generate electricity would be performing metabolism, and could do so autonomously if it were so programmed (though autonomous feeding is not a criterion for life).

4. Growing is simply building more of oneself autonomously. There's no reason to believe a robot could never construct and add additional subunits to itself. I can't think of any functional niches such an ability would be useful for, but that doesn't mean it's impossible.

7. Reproduction is simply making more of oneself. In the case of robots, it's typically more efficient to have a separate robot that handles all of the robot production than to include autonomous reproduction abilities in a robot that's also serving other functions, but being less efficient than alternatives doesn't mean it's impossible.

All of them are in fact possible. It would be inefficient to shoehorn a robot into fitting some of them, but that's just further justification for not considering robots to be somehow less than organic life ("you don't qualify as living because you have more efficient ways to carry out the tasks that we demand entities do in order to define them as living therefore we're going to deny you rights that we earn by being alive").

bulbinking posted...

Do some learning about neurobiology. The beain is waaaaaay more complicated than how you are describing it

I never actually said anything about complexity. Just the fundamental principles that are involved. Absolutely everything the brain does can be boiled down to electrochemical signals. The complexity of behaviour that arises from that stems from where, when, and how frequently those signals fire, and it's miraculous that such a fundamentally simple, almost binary system (an action potential either fires or it doesn't, with the refractory period setting it apart from something truly binary) can give rise to such a magnificent diversity of functions and behaviours. But fundamentally, even human sentience is just electrochemistry. Everything about who you are comes from sodium, potassium, and sometimes calcium ions crossing membranes. Why, therefore, could artificial sentience not similarly be them product of electricity? It'll be really bloody complicated, sure, but there's no fundamental reason to think it'll be impossible.

bulbinking posted...

Reductionism is not logic.

It can be. Your position is that artificial sentience will never be possible. Mine is that there's no reason to believe that because there are no fundamental differences between natural and artificial intelligence. That doesn't mean they're the same, but if it's possible to build up the complexity of natural sentience from simple electrochemistry, insisting that it's impossible to do the same for artificial sentience from electricity means you have to find a step in that process that will truly be impossible (as distinct from one which we haven't solved yet).
---

This is my signature. It exists to keep people from skipping the last line of my posts.