Print

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

[Girlinhat]

But then you're into the issue of practicality and design.  If a robot doesn't backup its brain, then is it different than if it did?  If I keep a journal is that any different than if I didn't?

[Neonivek]

But then you're into the issue of practicality and design.  If a robot doesn't backup its brain, then is it different than if it did?  If I keep a journal is that any different than if I didn't?

Then you have to seperate amnesia with death.

[GlyphGryph]

The thing is, humans are different from robots. We have no hard drive, no hardware backup. Hell, we don't even have ram.

We are basically  pure, 100% CPUs.

So assuming that computer AI die simply because they are shut off would be rather silly, I think. Or at least in the meaningful way. I guess it could be considered "dead" in the medical sense - but people have been dead and then gone on to live perfectly happy lives without being bogged down by philosophical questions, so I don't think it's a huge issue.

[rausm]

But then you're into the issue of practicality and design.  If a robot doesn't backup its brain, then is it different than if it did?  If I keep a journal is that any different than if I didn't?

Then you have to seperate amnesia with death.

Consciousness is impossible [to develop] without continuity. To go from A to B we have to remember where we came from and where we are going.

It's like talking about "what if water was dry". Either there is continuity as enabling force for consciousness or there isn't and consciousness won't develop. If you restore from backup, it is as if you went back in time, or never wiped anything in the first place.

Now, you can put someone/thing asleep and transfer them/it into different environment, but it would still be them, acting on their past experience.

Amnesia is mostly partial, and not affecting ie. muscle memory, speech etc. Not really a good example.

The journal is also a bad example, as to make any use of it you need to be able to read it at first (how, if that ability was wiped).

An always available, implicit/built in, large-enough "journal" (conscious+unconscious+muscle/etc+genetic memory in our terms) is, IMO, predisposition for sentience/self-awareness/conciousness.

RAM is also a bad example/analogy. RAM is just an implementation detail of a machine running the simulation. The simulation itself is not aware of it.

Also, just being able to survive, move or push a button to get food doesn't, IMO, imply consciousness/self-awareness. Just being aware of ones needs and environment doesn't make a being sentient. That just grants ability to survive. Awareness of oneself is the most important of the three, the other two just enable it.

Don't expect that "grandroids" will become self aware. Or thank you for taking care of them (unless they develop religion first; they are not complex enough, you computer is not powerful enough => it won't happen; there is no way to sensibly hard-code this into the simulation without destroying the "evolution" part of it).

[rausm]

Or in other words:

Jenny: Mommie, are we all just a simulation running on a big computer ?
Mom: No-one knows, baby-girl. Some might think so, others might call this computer God.
Jenny: And what if the computer was ever re-started ?
Mommy: We wouldn't even know. But one thing is for sure, you wouldn't have to wake up for school the next day. Now off you go to sleep.

 

[Rowanas]

I think we're too technologically primitive to create true artificial life just yet (at least, life that follows our definition of sentience) and we've got our heads jammed too far up our own arses to recognise sentient life if it didn't meet our conditions. I'm not even sure that self-perception is a necessary part of intelligence.

Closer to topic, though: He's hyping the "artificial intelligence" too much for me to buy in. If Alife had been conquered, it'd be bloody bigger news.

[Aklyon]

Well, when one works on Alife all the time (or as much as possible), you'd think thats what he would hype.

[rausm]

I think we're too technologically primitive to create true artificial life just yet (at least, life that follows our definition of sentience) and we've got our heads jammed too far up our own arses to recognise sentient life if it didn't meet our conditions.

There is nothing wrong with reading/watching shitload of sci-fi. But do not read too much into it.

Those presuming that "we wouldn't recognize sentient life because it wouldn't meet our conditions" don't have their head up their ass ? How so ?

I'm not even sure that self-perception is a necessary part of intelligence.

Intelligence != Consciousness (also Self-perception != Self-consciousness != Self-awareness != ... whole lotta other self-things).

A dog / rat may display intelligent behavior, but are they conscious ? Lets say intelligence is also one of the raw, enabling conditions.

Spech / communication (sufficiently sophisticated) would be yet another one. Let's say that for consciousness as we know it, there needs to be an above-critical amount of enabling conditions.

Better to talk/reason about something we know/can imagine than to say "fuck thinking, reality will be different" and start pipe-dreaming about "what if universe worked differently and we met a life-form so alien we cannot even imagine it... It would be different, somehow, like, i don't know but it would really be different".

Anybody interested might want to read up on experiments with apes (who are able to learn sign-language) and big parrots (who are able to talk, not just repeat but understand the meaning of ~4000 words, recognize them when they hear them and reply in context - in effect have conversation). Interesting question would be whether they discovered their concept of I (which they demonstrated) or whether humans imprinted it upon them through teaching them to communicate and therefore to think like them.

It would be interesting to select a big group of apes, train them individually (to not hold each other back / fall back to their natural behavior), then put them together and monitor the development of brain/brain activity in the generations to come.

Closer to topic, though: He's hyping the "artificial intelligence" too much for me to buy in. If Alife had been conquered, it'd be bloody bigger news.

He's not hyping (quite the contrary, he is very humble and is constantly saying that he doesn't know whether he'll succeed in creating something at least self-sufficient). But he's got further than anybody before. What he (partly, for the time being) got is something that, given sufficiently large storage/computing capacity, could develop into self-aware beings.

In - of course - very simplified environment; but hey, if it works, the simulation could be gradually "deepened", to the point where you then just replace the simulation with robotic body & set of sensors and ... (sorry for the obvious joke) get maimed because you are mistakenly interpreted as danger ;-))))))

By not trying to "create intelligence by stuffing more and more data into database hoping it will become self aware", but mimicking life as we know it, he might finally create something not requiring to be programmed, something able to "program itself", learn.

Once you have a dog that doesn't age, you can always teach him new tricks ;-)

[Mistercheif]

Why is electricity so different?

There is a difference between death and suspended animation.

When you turn off a robot, even a sentient robot, it isn't dead (ignoring that by all technical definitions it isn't alive either) it is just suspended.

That somewhat depends on how much RAM is being used.  RAM loses its memory when it loses power, so a reboot would wipe it.

Unless it backs it up or it has a saving function.

Or it uses that nifty new MRAM that has the nice convenience of being nonvolatile. Of course, there's the problem that it is nowhere near the memory density of volatile RAM, and the only capacity currently in commerical production is 64MB, but it still would result in an AI that would merely "go to sleep" when power is lost.

[Trollheiming]

The thing is, humans are different from robots. We have no hard drive, no hardware backup. Hell, we don't even have ram.

We are basically  pure, 100% CPUs.

We do have short-term memory and long-term memory. That's not like a CPU. A CPU doesn't have any storage at all, just accumulators that read in data from various RAM and ROM sources on the bus.

[GlyphGryph]

Of course CPUs have storage, don't be silly. Computers wouldn't work very well if they didn't.

And we basically use that to fake memory. Which is why we lose it when we use it, if you will. (Yep, our memories are erased every time they are recalled, what a strange brain we have!) It's a side of effect of using a processor as a memory storage unit. Well, a cluster of processors really. Once they "process" the memory, it's gone! *poof*

Of course, our brains are clever, and part of where they send the newly recovered memory is right back to the place it came from, so it ends up sitting there in the registers again ready to fire.

[lordcooper]

People are not computers.

[GlyphGryph]

Neither are Grandroids, for that matter, at least in the way you mean. But certain structures are certainly analogous.

And people can be computers.

[Trollheiming]

Of course CPUs have storage, don't be silly. Computers wouldn't work very well if they didn't.

Well, I don't know much about neurobiology, but I do know that CPUs are just a program counter and a set of opcodes. They boot to a point in memory and begin executing the code there. I've programmed some 8-bit microcontrollers before, pretty much identical to the first 8086 IBM architecture.

People are not computers.

1 + 1 = 2

Assertion falsified.

[PTTG??]

Computers and brains are both information processing machines which can express self-modifying behavior patterns. Nonetheless there are vast differences in scale, capacity, and operational mechanism. Everyone happy?