Headless Client (spring without GUI and Sound)

[zwzsg]

Sentience is unlikely to be duplicated in silicone

I agree.

Disagreed.

However, I'm fairly sure that once sentience is duplicated in silicon, people will argue for a century that it's not "real" sentience, they'll change sentience definition to something else or more vague to shield themselves from the idea that mankind might not be that special.

[lurker]

silicon

silicone

[Neddie]

I agree with your statements about what would happen if the mind were duplicated artificially entirely. I'm surprised, actually, to encounter somebody else at last with such a realistic and dismal view of human nature.

We're not special. However, I doubt, without near complete functional understanding of the human mind, we will ever be able to duplicate it in full. The naive neural complexity argument has never been reasonable either, a system is not merely the power it can bring to bear. Even if the connections could be modeled, they would then need to be tasked and used. The mind has an inherently different structure from any computing system we have devised, and none of the proponents of equivalent artificial sentience and electronic human modeling seem to take this into account.

It may be possible, but it seems unlikely that humans have the capability, and since we're the only beings capable of complex construction/intellectual synthesis in our own perception of the universe...

Of course, I'm sure this will never be accepted by the transhumanists who delusionally presume themselves capable of transcendence and perceive consciousness as a boundless construct. You know, the other group of people who think we are special.

[lurker]

Specific duplication of the human mind is overrated.

[LoidThanead]

Agreed with Lurker. Being an AI student, this topic is something we've philosophized about.
My personal view is similar to zwzsg. Even if we do manage to achieve AI that displays sentience, people will still argue that because humans made it like that, it's not truly sentient, or some argument similar to that. (Heck, we don't even have a precise definition of sentience. We really just asume that we are sentient ourselves, then try to figure out what the word really means. )


However, for most practical purposes it's really not all that useful to copy human intelligence. Humans are generalist. Besides their work (say, watching surveilance video screens,) they need to know how to walk, talk, think about what to have for dinner tonight... All that is useless baggage for an AI that has to recognize threats based on the video take. It should be a specialist.


Wow... how did this discussion get from headless clients to philosophy?

[AF]

Partially because that word sentience is misused, especially from startrek. It could be argued that apes and most mammals are sentient, that doesn't mean they're capable of starting a civilization, conversing in 10 languages, or building technology

[lurker]

Yeah, they just live in tribes, can handle a couple languages, and build technology. Wimps.

[smoth]

However, I'm fairly sure that once sentience is duplicated in silicon, people will argue for a century that it's not "real" sentience, they'll change sentience definition to something else or more vague to shield themselves from the idea that mankind might not be that special.

it's fine, cue the giant robots

[hoijui]

...back on topic again (headless client):
there are lots of classes under rts/Sim, that have draw code (Projectiles eg.).
If the simplest approach for me is to put #ifndef HEADLESS around everything that draws and sounds in the engine, and all includes related to that...
is there a better approach, with less then 2 * the time investment of this approach?

[imbaczek]

that's the idea, pretty much. also remove gl, sdl, x11 and openal from linked libs and the moment it compiles, it should work. you can compare sync by joining with the headless client and goofing around.

[hoijui]

ok

[hughperkins]

Hey! Cool to have some people around who feel similarly that computers will be strongly intelligent one day.

I feel similarly.

By the way, someone once said something like 'asking whether a computer can be intelligent is like asking whether a submarine can swim', and I feel there's a lot of truth in that.

Right, back to headless clients, going to read the rest of the threat now

[hughperkins]

Three years ago, one of the main cpu hogs on Spring was actually pathing rather than graphics. Even now, I'm not sure if that's not no longer the case?

Build 100 units, and if they sit there, everything will be fluid. Now tell them to go somewhere, and you get just a few frames a second? (at least: on an eeepc 4g, with a 600megahertz processor and 512meg of ram...)

[hoijui]

Build 100 units, and if they sit there, everything will be fluid. Now tell them to go somewhere, and you get just a few frames a second? (at least: on an eeepc 4g, with a 600megahertz processor and 512meg of ram...)

hehe :D
just a few frames with a 600MHz CPU...
hehe :D
... in brackets!!
:D that was a good one! :D

about AI getting strong..
yeah man! :D
i believe:
the question of whether a computer can be intelligent or not is actually the question of how to define intelligence combined with some other stuff, potentially.