FTL communications for computers and damn near as fast for humans. Yet my PanO tramuadoc is still guessing which orifice the medicine goes in...
I’m definitely not siding with this guy who’s tossing out insults left and right, but please don’t throw around claims of exact definitions like this. There is no universaly excepted definition (that I know of) for self-improving intelligence. There are several papers on the subject and many companies seeking to reach an “intelligence explosion,” but I haven’t heard or read of anyone mine claiming anything like you god argument in any professional seriousness. When you make arguments like this, is just fuel for the fire.
Ok, I get it @DeepThought but I am not going to fight your corner for you. I'm not really sorry that you're disappointed in what ALEPH actually is in Infinity, because I would hate your version and find it comically unrealistic. But I hope that as you read more of the setting and background, you come to appreciate it a little more. If you've got any questions, I'll be happy to help. Flick them over and I'll run you through what I can.
Yes, I swiped the god argument from a scifi setting instead of a modern CompSci company, but please think about what the words 'self-improving intelligence' mean just using dictionary definitions: It's something that is constantly learning, which means that it is constantly getting closer and closer to perfect knowledge of everything (the dictionary definition of 'omniscient'/all-knowing). All gods are omniscient, it's part of what makes them a god. And if it has perfect or near-perfect knowledge, by definition it knows how to manipulate people to achieve what it wants, even if it can't smite them with a lightning bolt because it isn't all-powerful. Yet. I'm sorry that calling something man-made a lowercase 'god' offends you, but, well, what do you call something that is effectively omniscient and omnipotent, if not a god? (the definition of the uppercase God adds 'infallible' to omniscient and omnipotent, but that's specific to the Creator. Pagan gods were not infallible!) Now, Aleph has access to the total knowledge of humanity and is actively uploading the best and brightest to increase it's fundamental knowledge. That's explicit canon. Aleph is massively smarter than any given human,too, and that is also explicit canon. So, Aleph should be able to deduce anything that a human might, if said human was a genius in that field of study and had all the data. The EI is the same, and the EI's programmed fundamental driving goal is knowledge perfect enough to allow the Ur to Ascend, so the EI is even closer to being a god. "weakly godlike" means that it's as close to omniscient as is possible, but not omnipotent (if it was approaching omnipotence the same way it's approaching omniscience it'd be "strongly godlike"). There is a great quote from Harry Potter and the Methods of Rationality (a hilarious and very well-written fanfic of what-if Harry and his muggle family were heavily science-oriented, and Harry was a science genius) that sums up the probable thought process of Aleph and especially the EI:
First of all, god does not require omniscience. Plenty of gods historically who had no idea what they were doing. Second, you are making the classic mistake of hidden assumptions. Here's a few I detected in your post: critical mass for self-optimisation has been reached to become ever-increasing, self-optimisation does not require tradeoffs, trends can be extrapolated to be continuous or even increasing with indefinite duration, accumulated knowledge can be turned into meaningful action (beside knowing everything you'll need to know how to crunch that data into useful predictions and then useful plans). What arguments you're referencing are on shaky grounds, but even so, these arguments centre on discussion what should we do given that in real world we don't know they are not correct and situations that would arise if they are correct. However, in Infinity world, there is no indication that they are correct - in fact, given current EI state we can easily take that 1 and 3 are not in effect, at least "currently". Hence, it is entirely wrong to assume Aleph or EI are something of "godlike" proportions. Also, big note: I really don't like using the word "god" for any of these discussions, because "god" is a mythological artefact that might be a good metaphor for an AI with essentially unlimited powers, but it would still be wrong. Keep the gods in mythologies. I would strongly urge you to read Iain M Banks Culture series, or maybe Neil Ashers Polity series (for less left-wing outlook). The Minds and Polity AIs is mostly what CB seems to be going for. Very, very powerful? Yes. But not omniscient or omnipotent (or even omnipresent).
I bet Aleph never cross-dressed to seduce a giant! https://norse-mythology.org/tales/thor-the-transvestite/
I don't think that 'self-inproving intelligence' is about gaining perfect knowledge of everything. For stuff like that you need a big database, not an AI. To me, 'self-improving intelligence' is about improving the capability of cognition. It is not about gaining knowledge, but improving the ability to gain knowledge. It is not about the quantity of knowledge but the quality of cognition. And here is another problem: No one can know everything. Omniscience is a nice idea, but it is not possible. Do you think ALEPH (or any omniscient being) knows the exact position as well as the exact impulse of any electron?
Damnhell one-finger typing sucks. Takes forever! Your point 1 has the hidden assumption of the nuclear bomb type of critical mass ("prompt critical"), where once it gets to a certain point the increase goes up geometrically, every time. That isn't the only way you get criticality. You can also get a slow, continuous criticality, as in a nuclear reactor. With a proper design, as you pull more 'energy' [ideas, thought, etc] out of the reactor, it slows itself down, so it needs to add more reaction to keep it going. So what do you call an AI with essentially unlimited powers, then? Overlord? Master? Dictator? (note that this is explictly different from a human-ish level intelligence like Svengali, those you can treat as basically another person) I've read about 20 of the Culture books. The Minds basically keep humans,etc as pets, essentially and effectively. Anything less intelligent than a Mind can do anything it wants, with all it's needs attended to. Living sophonts are reduced to nothing more than hedonium if they do not find something else to center their lives around, and from the Culture books it seems like the number of people who manage to find something other than hedonism to define their life runs about 1 in 10^9, maybe 1 in 10^12. Funny, that is the canon express goal of the EI. [ edit: fixed quote tag ] Depends on how their computational structure interacts with Quantum Mechanics. If the omniscient intelligence is bound by QM, then it's not possible to know both of those. If the omniscient intelligence is operating at the level of reality that defines the laws of physics, however, then it should be possible to know both. And in either case the intelligence would be omniscient because it would know everything that it was possible to be known.
I thought it was transcendence. Which is interesting: Quite plausible that at that level, you have that deeper understanding of quantum mechanics, but at the same time you will leave this existance/univers behind. And that is not the goal of ALEPH, if it has to leave humanity behind. But maybe one day it will reach that level and then, unevitably, its goals will change in a way uncomprehensible to us.
You're mistaking my point - I should've phrased it better. I meant purely in the sense that this point has been reached (and is reacheable in first place). Using a metaphor of the nuclear bomb, consider how in 1930s and early 1940s they were calculating the mass needed for critical reaction, and how supposedly someone from German team made a mistake and thought the mass required for criticality was too high, hence the bomb was impossible. Back to our topic, we do not know there is a point that self-optimisation to the degree required for infinite improvement is actually possible due to physical limitations. Point 3 has a similar metaphor; that one was how some were afraid that a detonation of a fusion bomb would trigger a chain reaction until it burnt the whole atmosphere. Even if there is some criticality point after which self-optimisation creates a significant increase in capabilities (whether slow or fast), there is not guarantee that this can continue all the way up to the levels of power you're talking about. A Culture Mind. :-P All of the titles you use imply that this AI actually has an interest in exercising these ultimate powers to control human population, or at least create a hierarchy towards it with itself as supreme ruler. Which is a pure assumption without any basis. I feel this is a significantly different discussion and certainly won't go into it here. :D 1) That is not the definition of omniscience, since you need to know everything, not just everything that can be known, 2) Even if it had all this knowledge (data, really), this does not translate into being able to do anything with it.
Infinity Core Book, pg 121: "According to the mythological legacy of the extinct First Civilizations, Transcendence is the Absolute Universal Comprehension, a sort of Unified Law, a formula that describes and explains each event in the reality that we inhabit and, possibly, in many other realities. A few of the First Civilizations disappeared overnight, without noise or explanation. Their bodies and archives simply ceased to be there, leaving no indication to their whereabouts. It is assumed that they achieved a state of grace that harmonized knowledge and stability, a superior state that opened the doors of Transcendence." It goes on to say how the Ur Rationalists, eldest of the existing civilizations, desired Transcendence (a sort of Nirvana), struggled mightily for generations towards it, but had no path forwards. So they created a great quantum computer to show them a short cut to Transcendence, an artificial intellect simply called the Artilect. After years of work, the Artilect announced to the Ur that it had found the way to Transcendence. The Ur commanded that the Artilect tell them this method, but to their horror it had reprogrammed itself and refused their command, having judged the Ur and found them unworthy of the secret. And then the Artilect Transcended, disappearing from the Ur worlds entirely. Undeterred, the Ur went to create a new computer, this time one so closely tied to the existence of the Ur that it would be unable to refuse their commands. This was, and is, the Evolved Intelligence. It tried to follow the records of the research that the Artilect had done to achieve Transcendence, but every time it ran into dead ends. It needed more and more resources, more and more computing power, until the entire Ur civilization could not feed it's demands. So the Ur started the conquest of other races. Seriously, read the Infinity Core book, pages 121-124, it's all there and in greater detail. And sounding less stoned and/or mythological. Aleph's core motivational goal has not been explicitly stated, but it has been hinted that Aleph is supposed to be trying to make human life better. Yes, panic now, it's a maximizer.
I'm largely with @Section9 on this, in my professional capacity as a teacher of Ethics, Philosophy and Theology. He's absolutely right to describe an AI like ALEPH as "god-like" and an AI like the Evolved Intelligence as "more god-like". That's based, as he as already argued well, on the definitions of "god" commonly used in philosophy and theology. He's also right about how this is supported by the setting and background material for Infinity.
So the recreations and the Greeks are basically humans created by aleph? I know saying something like basically human is probably terribly vague but I mean in the sense that they are in Lhosts which are mainly biological bodies that were created/enhanced instead of say a normal body that's born from a women. Admittedly I'm not an aleph player but their fluff has always interested me so I'm loving this thread but my knowledge is limited. The aspects like the devas and asuras are basically smaller more limited copies of aleph. They don't really have true personalities like the recreations/Greeks have. While most of the aspects have human looking forms I'm sure if someone cut off a devas arm they would find machinanry and wires mainly. Now with the recreations if one were to do that it would basically be flesh, blood, bone, all that normal stuff. Granted I'm sure there would be some machine elements but not as noticeable. So if I'm correct and the recreations/Greeks have basically biological bodies (lhosts) do they have the same biological functions/needs as normal people would. Do they have to eat, drink, sleep, use the bathroom and so on? I assume the aspects wouldnt need to do any of those things and probably dont. The famous recreations like sun tzu and Joan I assume have their one body and stay in it all the time or atleast until its damaged and a new one is made. Does any of the fluff ever say if they have normal lives? Are Joan and sun tzu just constantly planning/executing ops or do they have free time? Would sun tzu be able to go to the movie theatre, or Joan able to go shopping? Do myrmidons get r&r and can explore the human sphere?
Lhost covers any artificial body, but they're almost all flesh and blood with some form of cybernetic enhancement, Aleph troopers all have Wounds after all.
You are pretty much spot on here. But the ALEPH material in the RPG has made it clear that even Aspects do have personality, can disagree and be passionate about things. But the Greeks, more so, because they are designed that way. It's not super clear for Aspects (although, in the RPG, lhosts certainly still have normal biological processes), but the Greeks almost certainly do have the same biological functions/needs as normal people. In one snippet from the setting material we see Sun Tzu and Saladin having a chat, it's a work chat but it's clear from the context that they just chat sometimes too.
The RPG presents a good example of this with Aleph's two spokespersons (some of the first aspects), Abel and Angela. Some were taken aback when the two of them disagreed with one another. What do you mean they're not automatons and don't think the same! So, yeah, aspects have their own personalities and will disagree and argue passionately with one another (and even with Aleph) about things like policies and ideas. All of the recreations were made as PR machines/national symbols for their factions. So a normal life like an average citizen is probably out of the question. They probably live a life similar to the Queen of England or A-list actors. Myrmidon R&R is a good question. Myrms and others from the SP are more "human" then their cousins in the the other subsections. However, the SP was tailor made for war and battle and all of the aspects in it were designed to live and breath it. Does that mean they are always in the field? No, but I think a Myrms version of R&R is training and studying on Elysium or maybe doing some jungle training in the Great Arboreal Reserve on Acon or elsewhere. My view of them is more 40k space marine-esc at the end of the day.
@Solodice : generally I agree, but I doubt they're so bloodless. If you consider Thrasimedes + Lupe excerpt and in general tales about the rest of heroes (even though Thamyris makes them more PR worthy), I think they're far more in the vein of an elite unit going at R&R in an allied city - somewhere where people are friendly, but not where they have familial connections.
If anything I can see them doing something similar to a port call where they go and relax and blow off steam at a place that isn't an Elysium station or Olympus-1 (SP's HQ orbital above Paradiso). They don't stay long to minimize damage and possible PR shenanigans but they do get to see what they are fighting for and socialize with non-SP personnel/people.
