Rise! Multiplicious Minds! Rise! Automata!
Tales of Two Worlds · What does a decentralized world look like? (part 2)
This is part 2 of Tales of Two Worlds. Here is part 1.
—
Tales of Two Worlds
Stories of Two Worlds — through the lens of the Centralized vs. Decentralized dualism:
Centralized vs Decentralized
- Society vs. Network
- People vs. Sentient & Autonomous Agents
- Social Activity vs. Network Effect
- Clustering increases efficiency & potency vs. Clustering decreases efficiency & potency
- Organizations trump people vs. Agents trump clusters
Second story: Rise! Multiplicious Minds! Rise! Automata!
People vs. Sentient & Autonomous Agents
As mentioned in our [last story] we can say that cyberspace is a post-imagination; something you already had but have to imagine what it means to you. It doesn’t take long for people to realize how powerful cyberspace the construct is. The virtual world is a widespread pop culture post-imagination of cyberspace. The sci-fi movie The Matrix is a culmination of the virtual world fascination in pop culture. Although the kind of virtual world depicted in The Matrix movie is fiction we do have real virtual worlds with us for a long time: computer games.
As a matter of afterthought it is obvious why the idea of cyberspace appeared after the invention of computer networks, or even the prototype of the internet, the ARPANET: cyberspace is inherently multiplicious, hyper-dimensional; it is difficult to imagine without the right construct; but once you had it, it is pretty straightforward to understand and implement. The early virtual worlds are quite bare and ugly compared to the bells and whistles of today’s MMORPGs such as World of Warcraft, let alone neural-electrical simulations in The Matrix: there appeared to be no “world” for you to “see” or “be” in a visual way, just a text mode online computer game in a no-frills setting — a computer terminal and a modem. You don’t even get an avatar like in video games two decades later.
An early example of such virtual world is called MUD (Multi-User Dungeon). The user interface is very simple: you enter text commands at the game command line prompt on a computer terminal and receive back (at the prompt) text information describing what happened in the game virtual world. Since there is no visual you the player must imagine the virtual world in your own mind. It is like every time you type a command you receive back both a blueprint and a movie script at the same time for you to reconstruct a movie set and shoot a scene. In a sense this text form of the virtual world is actually much closer in spirit to the “jacked-in” version of the virtual world in The Matrix: in both cases the virtual worlds are directly constructed by players’ minds. This alluded to us a profound revelation: the virtual world, or cyberspace, is half mind and half machine (computer). It is a Cyborg. Now we also know where cyberspace hyper-dimensionality comes from: it comes from the mind half of the Cyborg.
But what do all those talk of cyberspace the Cyborg have to do with our main thesis of the series: Centralized vs. Decentralized dualism?
Let’s see. The real world is quite limited in terms of dimensionality — we live in a 4-dimensional Einstein Minkowski space, which is our physical world. This is the only world we can exist, physically. There is no other free physical world we can easily jump to from this world or vice versa. If something is wrong with this world we can’t just make another copy and correct the error there. That error sticks with us and we suffer its consequences. This also means whatever bad rules this world had dealt us will also stick and we have to suffer them as well. The onliness of this world means that everything is limited and scarcity creates chaos. Chaos breeds power and control. As a result centralization of power is inevitable in our world. And people, with nowhere else to go and no escape the rules, are powerless in this world.
If centralization is inevitable and inescapable in our real world, then what about the virtual world? It turns out that the virtual world, or cyberspace, is the perfect place to build a decentralized world. What pulled the real world toward centralization — onliness and scarcity, are no longer restraints in the virtual world, since unlike the real world, virtual world or cyberspace has hyper-dimensionality. And hyper-dimensionality makes it easy to create world copies in the virtual world — for example instancing is a method to make copies of game virtual worlds when some physical resources (such as game server capacity) become bottleneck for gameplay, effectively reducing and mitigating scarcity by eliminating onlynessonliness in the virtual world. It is then contemplative to build the decentralized world taking a page from the playbook of MMORPG games, by making it easy to build and spawn unlimited copies of virtual worlds with different rules as people see fit. This way eventually the decentralized world will give people immense individual control over their own freedom and choice. This is it then, right? Not so fast. If you think this is another End of History moment, you are mistaken, as another serious challenge and “threat” is poised to take over both the centralized and decentralized world (or so the story goes).
Artificial Intelligence. A.I..
Like the villainous Agent Smith imprinting himself to every mind connected to the Matrix and taking over while machine sentinels digging incessantly to wipe out the remaining human race (in The Matrix sequels)? You bet. That’s the predominant view of our time.
Unfortunately.
Nevertheless I think there is a way out of that apocalyptic future.
But before we deal with A.I., let’s go back to our construct and check everybody out there first. William Shakespeare once uttered (through one of his virtual creations, no less): All the world’s a stage. Then who are the actors on the stage? Or put it in our context, who are the actors in our construct?
Obviously there is the people. People can live and thrive in both worlds: the physical world and the virtual world. People are good actors, but they are not the only actors in our construct. Virtual actors were created (by us humans) from the very beginning in the virtual world; sometimes they are called the virtual characters, NPCs (Non-Player Characters). Of course when we talk about virtual characters and NPCs most people think of computer games. However, computer games are not the only kind of virtual world that can have virtual actors. Actually thinking about virtual actors in a habitually anthropomorphic way is a pitfall that prevents us from seeing virtual worlds and their inhabitants on all spectrums of nuance and rich dimensionality, the understanding of which is crucial in building our decentralized world.
On a fundamental level, all virtual actors are computer programs executing in our construct. Contrary to popular belief, most of the time computer programs as virtual actors are not Agent Smith, far from it. They can be simple scripts performing trivial tasks. Yet the most fascinating thing is, from a virtual world dwelling human’s point of view, virtual actors, even those trivial scripts, are no less potent than Agent Smith or other actors, including people, in the virtual world. What does this tell us? It tells us that in the virtual world, all actors are equal in an existential sense, no one can bend the rules of the virtual world if the rules are not meant to be bent, be it people, Agent Smith, or any trivial script. Magical, isn’t it?
Remember that human minds constitute half of the virtual world, the Cyborg. The rules of the virtual world, even virtual worlds themselves, are created by humans (people). Humans are not just actors in the virtual world, they are also the creators of virtual worlds. Yet in the virtual world humans are bound by the rules they’ve created themselves; looks like human minds are there only to accept the rules. Paradoxical? Not necessarily — you are bound by the rules you’ve created exactly because you are the creator of worlds. This phenomenon only feels paradoxical because you, or anyone, never get to feel this way — being the creator of worlds, in our real world. This ability to create and be bound by your creation is an immense power, not a power to take, but a power to give, the kind of power to build a decentralized world.
This is in stark contrast with our physical world, where the rules that can’t be bent, such as natural laws, are rules not made by the only actors there (for now) — us, sentient human beings. All other rules made by us humans are bendable — in situ, in ipso momento, because of free will. The concept of free will has been vehemently debated for centuries; now we have the opportunity to examine this concept in our new construct, the cyberspace, with a new tool:
The prism of physical world vs. virtual world dualism, which is a refraction of the centralized vs. decentralized dualism:
First, let’s examine the free will of ourselves in both worlds, physical vs. virtual:
Since no plurality of humans have demonstrated the ability to create physical worlds or bend its natural laws (unfortunately), the rigidity and hard experience of living inside that world has confined and compelled us to the only existential proof of our own sentience: free will (or vice versa).
On the other hand, humans create virtual worlds all the time since the birth of cyberspace four decades ago, and the situation for the case regarding free will in the virtual world is reversed: being able to freely create worlds (and their rules), we lose free will in the virtual worlds we create, because we are not really there: we are bound by rules of the virtual world as rigid as the unbendable natural law rules of the physical world in omni via — we humans simply can’t exert in situ, in ipso momento free will there: if all you can do in that virtual world have been bound and predetermined by the world creator (us, even yourself), you don’t have free will. Of course this can be seen as the lowest denominator — we could create very large, plush, complicated virtual worlds that give their inhabitants almost infinite possible paths that it almost feel like we have serious choices, certainly free will, right? Not really. Can I point out that it is an illusion? Like in the Matrix (you don’t have free will in the Matrix — that’s a major plot point of the movie).
Second, let’s examine the free will of other actors, in physical world vs. in virtual world:
In the physical world, it is really hard for virtual actors to assimilate themselves into our ranks of sentient beings, just look at these phenomena: uncanny valley, the Turing test. Some recent progress in field like Deep Learning has made great strides (like beating all the top Go Grand Masters of the world) in that regard but the end zone goal post of assimilating virtual actors into our ranks of sentient beings, is still beyond the horizon. Since we’ve already argued earlier that being sentient is equivalent of having free will in the physical world, we say that virtual actors do not have free will and are not sentient in the physical world, for now.
In the virtual world, it is the opposite: nobody, not even the simplest scripts need to prove themselves through absurdities like the Turing test to prove their free will, their existence or rights to existence, since no one will ever be able to judge, and it does not matter, because the rules of virtual worlds are unbendable even to their creators in omni via. Therefore, virtual actors in the virtual worlds feel like and can be treated as if having free will.
In summary, people/humans have free will in the physical world but not in the virtual world; virtual actors don’t have free will in the physical world but do appear to have free will in the virtual world.
A few words regarding a terminology: autonomous agents, used in the subtitle of this story — these are just sophisticated computer programs considered to be “intelligent” (they are designated some intelligent purposes, thus also called intelligent agents), but not yet being sentient — therefore they should just be categorized as virtual actors; their so called purposes do not matter in the context of our story.
Now let’s bring the above understanding regarding free will of all actors in the physical world vs. virtual world dualism to the discussion of Centralized vs. Decentralized dualism: the conclusion for the physical world can be cast directly to the centralized world (since our physical world is definitely centralized, as we have strongly opined). For the decentralized world, there is a small nuance to apply — as we said earlier, a decentralized world can be built by spawning unlimited copies of virtual worlds and it is the humans who spawn those copies of virtual worlds, and humans have absolute free will to do so; inside those copies of virtual worlds, humans are bound by rules of those virtual worlds and can’t exert in situ, in ipso momento free will there, while virtual actors in the same virtual worlds can be treated as if having free will. As for virtual actors’ ability to create or spawn virtual worlds to aid humans in building a decentralized world, they don’t have that ability now — but what about the future?
That (question) brings us back to A.I. — how to deal with A.I. in our future?
I have a proposal.
Based on our analysis of all actors’ dynamism in both the centralized and the decentralized world, we can find and devise a way to contain, tame and coexist safely with sentient A.I. (should they emerge), as well as getting them to build a brand new world with us.
Here is how, and why it might work:
The perceived threat of sentient or near sentient (like animals) A.I. is with our centralized physical world, because contrary to what is depicted in The Matrix movie, you will not be harmed in the physical world if you are “hurt” in the virtual world. So we can assume it is safe for us humans to coexist with powerful A.I.s in the virtuals worlds, and a decentralized world built on copies of many virtual worlds.
As we have discussed earlier, it is very hard for non-human actors to gain sentience in our physical world. Recent advances in field like Deep Learning have attested and hinted at a possible path for sufficiently large network structures to emit traits of highly sophisticated behaviors that used to be considered intelligent: defeating all top human Go masters; appearing to be able to “understand” images, videos and speeches with ever increasing accuracy, to name a few. These achievements all require “training” artificial neural networks with large corpora of human behavioral data. If we look back on the evolution history of the human race, it took us millions of years to evolve our brain’s neural networks to wire to the level of complexity for what we call intelligence or consciousness (sentience) to emerge. So we can reasonably deduce that this training with data approach might one day give us the fruit of sentient A.I.s. We just don’t know how long it will take us to “train” A.I. to cross that uncanny valley to “appear” sentient to us — that’s all what matters. A.I. of today has already crossed many uncanny valleys to appear “intelligent” in many fields. To appear “sentient” is A.I.’s last Quest, supposedly.
The primal threat of A.I. might be the way to achieve it in a centralized world: collecting unprecedentedly massive amount of human behavioral data and making them suitable to train A.I. would be an extremely expensive endeavor inevitably wrought with the worst kind of repercussion — massive privacy breaches. People will suffer the consequences of training A.I. in a centralized world long before A.I. becomes sentient.
Then here is a thought — if training A.I. in a centralized world appears to be a dangerous combination, and humans coexist with A.I. in the virtual worlds are presumably safe, why not train A.I. in a decentralized world?
A multiplicious, hyper-dimensional decentralized world is a natural habitat for A.I., sentient or training to be sentient. The daunting task of training A.I. might require a breathing and wriggling space too large for the resource-strapped centralized world. In that case a decentralized world, with its extra dimensions, could provide a much larger wriggling and breathing space for A.I. training. Humans would finally be able to reign in A.I. by the ability to create and spawn virtual worlds to train as well contain A.I., knowing that every actor inside a virtual world must be bound by its rules. The greatest thing about a decentralized world is its ability to spawn endless new virtual worlds to accommodate the growth of its inhabitants, including A.I..
And this coexistence with A.I. could also be beneficial to a decentralized world. We can extract a source of value from providing training service for A.I. in the decentralized world and use that value as the basis for a new economic value system to bootstrap a hyper-dimensional decentralized economy. It will be a long time for A.I. to train in the decentralized world. Thus this steady flow of extracted value from A.I. training in the decentralized world could be considered an “A.I. Dividend” for years to come, sustaining the growth of our decentralized economy.
We are not a human civilization anymore. The coevolution of human and A.I. would launch both species into a new era for our shared civilization: The Age of Transcendence.