#On Artificial Life

1 messages · Page 1 of 1 (latest)

teal crest
#

So genuine question because I've now been thinking about this for a bit (Like 20 minutes, at best):

Our world has biologic lifeforms that are not necessary sentient. Yet, despite that, they are still considered to be living. Thus, with that and considering the fact that most people would not consider Neuro to be sentient (I think?), would she technically still be considered an artificial life, and therefore a type of lifeform?

(That's what thinking on empty stomach gives you, folks...)

cerulean field
#

There is a certain level of complexity I would think to count as living, like how a virus isn't actually alive

#

"cant keep themselves in a stable state, they don't grow, and they can't make their own energy" Google on why biologists don't believe viruses count as alive

teal crest
#

So I looked it up (and thank you for this by the way, because it was interesting). I saw a classroom post from Khan Academy that explains that in order for something to be considered alive, it needs to have the following characteristics (That will translate for the purposes of this conversation so that it works with AIs, as a concept):

#
  1. "Living things must maintain homeostasis" (They need to be able to control their own internal state). I suppose that in this context, you could interpret this as an AI being able to self regulate it's own model and backend. 2. "Living things have different levels of organization" (A structure of smaller components that lead to a bigger whole) I supposed that this could refer to the different components of a AI model and how these can all lead to the model. 3. Living things reproduce. I supposed this would be regarding whether an AI can actively (Off its own volution) recreate itself, iterate, or create different versions of itself. 4. Living things grow. (Using energy and whatnot to develop) Here, since a machine does not behave like an organic being, I would interpret this as the ability to use resources (either be it, energy or information), to grow. 5. Living things use energy. (Here the text describes it as the usage of host-produced energy to provoke all the aforementioned growth) 6. Living things respond to stimuli. (Pretty self-explanatory, I think). Can an AI actively react to its environment (Host system, or other devices to which it is given access)? 7. Living things adapt to their environment. (Again, pretty self-explanatory) Can an AI adapt to its environment?
#

And sorry for the long write up, just thought that it was interesting to look into all of that.

#

Anyways, based on the above, I guess that neuro would fit within some of those criteria seeing as from my understanding (A very layman one, since I am no programmer), but would fail a lot of the other such as: 1, 3, and 5.

#

1, because I don't think that she can regulate her own prompt and personality base (In the same way that Vedal can); 3, because she cannot replicate herself (though it would be interesting to see if through the addition of new sub-components an AI could learn to create other AIs); and 4 & 5, mainly because I am completly unsure of how to categorize the generation and usage of energy in the context of an AI. We could interpret it literally as electrical energy but thus, it would mean that any future artificial existence even with the best of models (and sentience, if that is at all even possible), would not be considered as living since technically, it would still require (I assume), a load of external power to keep them running.

teal crest
cerulean field
#

I suppose in order to possibly count as living, it would need the ability to manipulate it's code significantly, and you could argue that Neuro generates her own energy by bringing in money to power upgrades and the electricity bill

#

And yea this is quite a interesting topic

#

So 4 and 5's energy could be translated as generating revenue from streaming in this specific case

#

But at the end, in order to truly live, an AI needs to know how to code itself

#

Well, and be able to

teal crest
#

Yeah, pretty much. I remember reading somewhere that Vedal did try to use Neuro as programming Co-pilot, but said that she wasn't great. Honestly, I would really be interested to see a livestream where Neuro is given access to a VM with MS Code project can given tasks to 'teach' her how to code, with maybe also some limited access to certain websites to search solutions like a human would. It would be interesting if through this she would be able to eventually create a rudimentary LM. From there, maybe, Vedal could also try to give her access to parts of her own code to see what happens (With a backup ofc, in case something goes wrong)

teal crest
#

Also, it would be interesting if there was a way to know whether this is an autonomous survival process that neuro is aware of, and therefore volunterily (At least, as much as she can, since technically she doesn't have a choice), partakes in because she knows that it is beneficial to her.

#

Honestly, it would be interesting to see/ test for week, if given given the ability to connect on Discord/ hibernate for a specific amount of time/ or do something else, if a model like neuro (with the necessary components added to the mix to handle all of that), if Neuro would activelly make connect everyday on Discord to talk (If she was told that her survival depended on that)

#

Man, it really is fun to be a backseater: propose a bunch of ideas that you don't know the feasability of, and hope hope that someone tries them lol

somber loom
#

Ignore the haters. To your point, despite the efforts of mankind, we still cannot define what life is... Much less any subcategories that may or may not exist.

teal crest