Bay 12 Games Forum

Wait, we talkin' UD, TFTD, or APOC? Or the Gawd-offal will-not-be-named one?

DF has two different definitions of 2D and 3D. The old school 2D vs. 3D meant back when DF only had the one z-level, Toady later added '3D' or multiple z-levels. The current 2D vs. 3D is about what method is used to draw the screen. With 2D, it's the CPU that's doing all the graphical calculations and having to share it's processing power with both drawing, and things like path finding. 3D shunts all the graphical stuff off to your graphics card, leaving the CPU to do the logic side of things without being dragged down with all that "Where the hell does this pixel go?" BS.

The first AI's with any real depth will start on the internet.

I guarentee it.

No other place is more suited to interaction, for a machine.

Could already have happened, at least the beginning stages. I mean, with all the adapting, self-replicating viruses (http://en.wikipedia.org/wiki/Plural_form_of_words_ending_in_-us#Use_of_the_form_virii) it's reasonable to suggest that there is at least life (In a digital sense) if not some early, rudimentary form of intelligence.

@Bauglir
Got kinda caught up in my current fortress, but I'm still popping back here now and then and considering your angle. I'm not the most intelligent being in existence, so I have to work through this as my tragically damaged brain (and ADHD) allows!

Corbald: Crashing *immediately* is generally considered a better idea.

Fortunately, I'm not having any issues/crashes that can't be found on the bug tracker. Oh, and great job Toady/Baughn/everyone else associated with this game. I'M SO PROUD OF YOU GUYS!!! 

Using the binary posted somewhere earlier in this thread (the win32 one that skips texture checks, iirc) I have been able to get ogl mode (VBO, specifically) working. Been playing around for a bit, and (bugs aside) I gotta say I like the new stuff. Burrows are FRIGGIN AWESOME!! and the new military setup, while confusing, (and buggy) is also exceedingly powerful. I haven't done much with the Hospital stuff yet. Also loving the new HFS...


I'm planning on creating a very large open area a few z-levels above the hot HFS for tree farming, but I'm running into the same issue with mud washing away that others are, so we'll have to see how that goes.

Only two more things that would make my fortress perfect. A rotation sensor, to flip switches when an axel/gearbox changes state, and an 'Alert Bell' that would toggle between two Alerts when a switch or pressure sensor changes states. (Hum... think I'll post those in the suggestions forum as well)

Bugs aside, I love the new content, and can't wait for my first siege to test out the military, and all this steel armor/weapons I have made.

Found that quote, finally. The first part sounds like double-talk, but is just setting some logical ground rules.. essentially self-proofs.

Axioms: A statement that there is no truth, if true, is false.  Nor can anyone testify that he has perceived that all his perceptions are illusions.  Nor can anyone be aware that he has no awareness.  Nor can he identify the fact that there are no facts and that objects have no identities.  And if he says vents arise from no causes and lead to no conclusions, he can neither give cause for saying so nor will this necessarily lead to any conclusion.  And if he denies that he has volition, then such a denial was issued unwillingly, and this testifies that he himself has no such belief.

Undeniably, then, there are volitional acts, and volitional beings who perform them.

A volitional being selects both means and goals.  Selecting a goal implies that it ought be done.  Selecting a means that defeats the goal at which it aims is self-defeating; whatever cannot be done ought not be done.  Self-destruction frustrates all aims, all ends, all purposes.  Therefore self-destruction ought not be sought.

The act of selecting means and goals is itself volitional.  Since at least some ends and goals ought not be selected (e.g., the self-defeating, self-destructive kind), the volitional being cannot conclude, from the mere fact that a goal is desired, that it therefore ought to be sought.

Since subjective standards can be changed by the volition of the one selecting them, by definition, they cannot be used as standards.  Only standards which cannot be changed by the volition can serve as standards to assess when such changes ought be made.

Therefore ends and means must be assessed independently of the subjectivity of the actor; an objective standard of some kind must be employed.  An objective standard of any kind implies at the very least that the actor apply the same rule to himself that he applies to others.

And since no self-destruction ought be willed, neither can destruction at the hands of others; therefore none ought be willed against others; therefore no destructive acts, murder, piracy, theft, and so on, ought be willed or ought be done.  All other moral rules can be deduced from this foundation.

"The Golden Transcendence" - John C. Wright

I think this pretty much covers what I'm trying to say.. If not, I'm open to more discussion

EDIT: Of course, I'm banking on the idea that a machine intelligence would be physically unable, or unwilling, to act in a way which it deems to be illogical (Data/Spock). Unlike Humans, which regularly act illogically. Morality can be Universal... Sticking to it may NOT be .

First off, number 2 is fatally flawed. Morality is not universal. It can be made 'universal' with logic within the human species, but interspecies relations and relations within other species may follow entirely different rules. For example, is the slaughter of thousands of fish more or less wrong than the death of a human? If, as I suspect you would, you answered killing the human is more wrong because humans are intelligent, who is to say an AI would not say the same of us. That the death of thousands of humans is less wrong than the death of one AI with superhuman intelligence. Morality is an evolved trait which is used to more or less keep social structures together. A non-social species, if it were to evolve to intelligence*, may not even acknowledge morality as a concept. This is why if we create super-intelligent AI, they must be started from a template similar to humans; to think of us as being essentially the same as them; part of their society. When you have an us and a them, it will rapidly devolve into us vs them.

Secondly, organic computers are not magical. They can not fit infinite numbers of transistors into a finite space. The reason organic processors such as the brain are currently more efficient than your PC's CPU is because they are massively parallel. A neuron maxxes out at around 1 or 2 kHz, whereas even a crappy PC goes at almost 2.5gHz. The difference is you have only a small handful of processors (1 to 16 or so) whereas you have billions of neurons. However, when talking of organic processing methods, what is being discussed (to the best of my knowledge) is essentially chemical reaction based. What matters is how they perform, not whether they are arbitrarily classified as organic or inorganic. In that regard, I'm fairly sure nanotech beats out organic processing (or likely will in a relatively short period of time) in terms of both speed and durability.

*speaking purely in hypotheticals here; social structure may or may not actually be one of the keys to evolving higher intelligence

As for your response to #2, I, respectfully, disagree. Thousands of fish, or a Human? Let's say that I don't choose to not make that choice at all, and (impossibly) physics forces me to. Would you agree to the 'Lesser of Two Evils' method? But also agree that BOTH are inherently 'Wrong?' I, am, of course, NOT a hyper intelligent AI, and I'm sure any answer I give could be picked apart by someone, but isn't that the whole issue in the first place? The pursuit of knowledge to help us to avoid situations like this? If, for example, I knew that one of those fish carried a gene that would one day lead to the development of sapience in fish, I might choose the human. Otherwise, I would choose the creature that already has volition, human or otherwise. That being which is most capable of self improvement/information processing, within the limited remaining life span of the universe. 1) attempt to find a solution in which no damage is done. 2) Failing that, minimize damage. 3) failing that, damage that which is repairable/replaceable. 4) Failing that, realize that since matter/energy is interchangeable, all that is really being lost is the information concerning the structure of fish/human.

Basically, you aren't disproving my point, just forcing me to do a cost/benefit analysis. Possibly some damage control.

I'm not saying that a hyper intellect is going to be perfect, but that morality IS universal. The big issue here isn't that statement, it's the fact that we exist inside said universe, and can't know everything. If we could know everything in the universe, we would have to exist outside it, which implies that there is more universe to exist in!


Point the Second. Sorry, I failed to point out in my OP that the 'infinite transistors' was a poke at so called 'quantum processors,' NOT the organic processors (read 'Carbon based Transistors.' Carbon is extremely resistant to heat, silicon is... well... not) ma bad heh. And I wasn't refering to the brain when I said organic processors, but rather new transistors based on organic molecules, rather than the good old silicon based ones we have now. Imagine the power of a microchip where each transistor was a molecule with only like 10-15 atoms in it. Even a huge molecule, say, the size of DNA, would be a major upgrade to what we have now. As it is now, Moore's law will eventually break down, due to that fact that we can only make a silicon transistor so small before it's own action burns it up, and I'm not even gonna get into Planck's constant here, as I tend to be long winded as it is!

EDIT: Wanted to also inject that morality is NOT a concept created to ensure social stability, but rather an analysis of least-loss physical principles. I'm trying REALLY hard right now to find a (rather wordy) quote that EXACTLY sums up what I'm trying (failing) to say...

EDIT 2: Can't find the quote    Read the trilogy: "The Golden Age" by John C. Wright. Specifically, book 3, "The Golden Transcendent" After Phaethon meets his father at the Solar Orbital Array, while preparing to enter the Sun, the group discuses Objective Morality. Phaethon presents the most exacting definition of the subject I have ever seen. If someone owns the book and can post the (admittedly multi-page) quote, please do.

Ok, I have been lurking here for some time, but I have to throw my 2c in here.
I make some logical assumptions, because without them, the entire world kinda falls apart anyway. These are some of the assumptions that science can't prove, but relies on just to exist!

1) The Universe is self consistent. If the rules for one part of reality don't accurately reflect the rules somewhere else, it's because we don't understand the underlying rules. (Black holes work, reality outside them works, so some super-set of rules is governing both)

2) Morality is universal. OR Two entirely logical beings, given identical information, will arrive at the same conclusion.

3) The basic law of economics reflects the basic laws of energy conservation.

4) Emotion (Drive) can evolve from an entirely logical system. Why not? Reality is entirely logical, but it creates emergent systems that SEEM chaotic. Such as the human brain.

We'll start with #3. AI won't kill us. it may upgrade us, though that would likely be an entirely voluntary situation. I say this because killing us would be counter-productive to it's goals, regardless of what those goals are! Think about it, if you have a job to do, why kill/destroy/remove something that can do other, less critical jobs, while you focus on the important stuff? Such as maintaining your systems, or some simple number crunching so you can free your own processors for other algorithms.

And now to #2. The more learned, and the more intelligent a human becomes, the less violent they tend to be. Look at any High School for an obvious example of this. Geeks vs. Jocks (sorry to any Smart Jocks here   ) This trend would obviously continue to higher levels as well. It's simply wasteful  to murder/steal/destroy. Even medical science is trying it's best to remove the cutting from surgery, for obvious reasons.

...to argue that moral judgements can be rationally defensible, true or false, that there are rational procedural tests for identifying morally impermissible actions, or that moral values exist independently of the feeling-states of individuals at particular times.

R W Hepburn
sums it up succinctly. Though I would go beyond that and state that the feelings of the individual(s) are an inherent part of the interdependent rule set that would be used to determine moral/amoral. (Two people want something, and have equal right to it, the person that feels the most strongly about the object in question has the most right to it. As a rough (very rough) example)

And #1 is self-explanatory. It's the basic concepts of science and relativity. The universe works because it makes sense, and it makes sense because it works. It's 'Self-Consistent.' Any part of reality will make sense if you understand all of the rules, and all of the rules can be derived from any sub-set of them. (Science is based on that principle, and I think science is pretty much a proven idea so far )

#4 MUST be true or AI will NEVER happen. If an entirely logical system determines that it is pointless, it will self terminate, to prevent waste. Self termination would be wasteful, and would circumvent volition (personal choice) and is thus illogical. It's one of the basic tenants of logic. Continues existence is necessary for volition and thus survival is the ultimate goal of any being. That right there is a 'Drive.' All other emotions/drives can be deduced from that source. (Freud was on to something here. Sex is the most important part of our psyche, only because of this basic rule, and we can't live forever, so we attempt to make the species (and our own DNA) live forever.)

So... Now that we have proven that AI will be our friends. Let's talk about whether AI can even exist.
AI = Artificial Intelligence.

Questions that MUST be answered to determine if something is AI:

What is Artificial?
What is Intelligence?

Sorry folks. Both of those questions are unanswerable. Arguments have be raging since Mankind could ask those questions as to what is this 'Thinking?' What does it entail? You can go look it up if you want. You'll find an answer, I'm sure. And it'll be wrong in 3 months (or 3 minutes). The definitions of 'Alive,' 'Thinking' and 'Intelligent' change all the time. Let's just call it 'The ability to use logical deduction, and process information, at at least the level that Humanity does (on average).'
Same with Artificial. EVERYTHING is natural, as it's made from/with the substance of our universe. Humans are nothing more that extremely complicate chemical machines. Are we 'Natural?' or 'Artificial?' Let's use 'Anything made by Humans, that is not, it's self, inherently Human. (Babies   )'

Can we create AI? Sorry to the doubters, but DUH!! If the universe can do it, then so can we (remember the self consistent universe bit?) Just remember, the difference between 'Organic' and 'Mechanical' is a VERY thin line, and entirely a matter of classification, NOT application. We are already doing research that says that micro-processors would work better if we used organic chemicals, instead of silicon transistors. (to the power of 10, if not more) It may even be possible to fit ALL THE TRANSISTORS YOU WANT INTO THE SAME SPACE!

In short (too late) Technological Singularity (The super AI version, not the people = computers kind) is;
A) Possible, and probably inevitable, if we can keep up our current levels of discovery.
B) Likely not a 'Bad Thing.'
C) Coming soon....