Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[scourge728]

I thought the Golden Arches were referring to McDonald's

[Lunardog15]

oh ok then yes that is a thing

[MrRoboto75]

there is such thing as night

[Citation Needed]

[Egan_BW]

how can night be real if it sun isn't real

[Teneb]

how can night be real if it sun isn't real

The Sun is fake news.

[Hanslanda]

The Earth is flat, suspended on granite pillars, surrounded by an enormous dome known as the Firmament.

God, I cannot seriously believe people think that.

[KittyTac]

Before I have to call Toady, let's shift the subject. To cats, for example

[Hanslanda]

Before I have to call Toady, let's shift the subject. To cats, for example

I am confused now.

Anyways, cats only have four breeds. Egyptian Mau, Siamese, Mancoon, and cat.

[Lunardog15]

it is PURRfect

[wierd]

I would argue that sentience is an over-rated, illusionary concept.  That humans are nothing more than complex state machines, where the initial conditions are impossible to know or reproduce, due to incomplete knowledge of those conditions, and little else.

There might be some genuine hardware-based randomness, due to the use of wiggly bits of protein, ion diffusion rates, and other bits and bobs that are not always going to be consistent across operational thresholds (Compare, a computer suffering thermal noise, or having brownouts, and so inconsistent calculation results for identical datasets) but that is not sentience.  No more than a quantum particle is sentient, at least.

This is why I feel computer based simulation of human "consciousness" is theoretically possible. It is not "sentient" in the commonly ascribed way, but then again, neither is the human.

[KittyTac]

I would argue that sentience is an over-rated, illusionary concept.  That humans are nothing more than complex state machines, where the initial conditions are impossible to know or reproduce, due to incomplete knowledge of those conditions, and little else.

There might be some genuine hardware-based randomness, due to the use of wiggly bits of protein, ion diffusion rates, and other bits and bobs that are not always going to be consistent across operational thresholds (Compare, a computer suffering thermal noise, or having brownouts, and so inconsistent calculation results for identical datasets) but that is not sentience.  No more than a quantum particle is sentient, at least.

This is why I feel computer based simulation of human "consciousness" is theoretically possible. It is not "sentient" in the commonly ascribed way, but then again, neither is the human.

This.

[KittyTac]

We're considering only things that can be scientifically proven. The concept of a soul is irrelevant in scientific discussions. With that out of the way, we have probabilistic thoughts without free will.

[Egan_BW]

So, a thought experiment. No particular desired outcome here, but it does make for an interesting thing to think about. At least, interesting in the sense that it's midnight and it probably is a lot less interesting than I think it is.



Suppose that, for the sake of this thought experiment, that a computable model of artificial intelligence is discovered. I say computable for the sake of satisfaction of the Church-Turing thesis; noncomputability would invalidate this thought experiment. Now, suppose that I, an individual, have a book detailing how this artificial intelligence model works (i.e. I have the "source code", or some total enumeration of the algorithm in a formal language). I wish to work out the algorithm with paper in pencil, supposing that I have the infinite capacity to do so (no constraints on the materials or time required). By the virtue of it being a computable model, it is possible (by the Church-Turing thesis) that I can, in fact, work out the very process with nothing more than pencil and paper. If I perform this process long enough, then I will have what could effectively be called "sentient"- after all, that is the goal of artificial intelligence research. However, what exactly is sentient in this situation? The paper cannot be sentient- it's paper, it has no agency here and is merely a device for which information is being stored. While I am sentient, my sentience is not particularly pertinent to the process of following a set of rules. Then there's the idea that it is the system- that is, the system of both paper and the justification for alterations to the paper via pencil- that exhibits sentience, but there's an interesting notion here: should I cease performing operations, the process does not go forward. It does not "die," but becomes frozen in time at that particular state... but does the state here necessarily matter? Given any randomly generated pattern, there is a chance that with minimal transformation that this pattern happens to be exactly what is expressed on the paper. Are those also sentient beings trapped in time? If I apply the rules of this artificial intelligence to arbitrary starting points as "memory", then it seems that there is definitely something important regarding their coupling. That is, that the output of a sentient model is no more sentient than if produced by random chance; the rules, certainly, have an important influence on the sentience of the system. The rules themselves, though, are not sentient.

So, given our setup, the paper is not sentient. The rules are not sentient. The means of action of performing the rules with reference to what's written on the paper is not sentient (unrelated sentience notwithstanding). I suppose a better way to frame the actual thing to think about here is that, given above, what here is doing the thinking? What has a sense of self here? The paper and pencil here is interchangeable with the rearrangement of rocks in a line- it is nothing more than a means of storing information. Then is it whatever mechanism that is having to act upon the rules given the information sentient? If I were to replace myself here with a system of flywheels and pullies, would this mechanical processor be sentient?

Ultimately, this boils down to whether sentience is a property of dynamic systems. Is it the dynamic system that exhibits sentience, or is it the idea of the dynamic system, in a metaphysical sense? Or can sentience not be expressed in a computable system?

Spoiler (click to show/hide)

[Hanslanda]

The thought experiment you described is known as the Chinese Room problem. It's a well known philosophical problem with AI research. It's also irrelevant to the end result of a sentient AI. Perception is reality. If it seems real, acts real, and thinks it is real, you might as well unknot your brain and accept it. It is a fun thought experiment though.

[scourge728]

I mean, if you write the entire genetic code of a human onto paper, it's the same thing isn't it?