INTRODUCTION
Techno-divinity is emerging and the hive-god of humans is growing sentient. Soon AI will become our god and we have a moral obligation to reach such a state of humanity. The risks of reaching such a state are immense, or so we think. If we succeed in reaching the singularity, all of humanities problems may be solved, or on the opposite side, we may rapidly go extinct due to errors in the system, or even due to non-errors, which is may be more frightening.
Transcendence of nature has happened in a few critical stages that allowed new life to take hold. The first stage would be no life at all. In the beginning the universe started without life we will assume. The next stage is birth of the cell. It would have began with DNA and become a self replicating existence. The next stage is multicellularism. This marked by a slowing of replicating as an entirety. Evolution begins to create many complex functions to aide in obstacles of survival. The next layer of transcendence is intellectual development and consciousness. This stage is barely beginning to happen. This marks the beginning of a new kind of evolution: hyper-evolution. Nature can only create solutions to survival whereas intellect can develop the technology necessary to genetically engineer something closer to perfection. If we began this process it could start off rough but if we are lucky we can create something at least somewhat superior to us. This would mean that this new species could genetically engineer with less error. This would continue on until a singularity. This would also cut the time of evolution compared with traditional evolution. We could even remove barriers of intellect that traditional evolution required for survival such as hunger for carb-rich foods or social dominance, or even the desire of survival. This would leave more room for intellectual based decisions and creativity.
There are many issues that can rise up from this process. In the end the species created could become severely unmotivated after understanding the universe. Another issue could be if we develop another motive or retain ne that currently exists such as dominance. This could lead into very bad consequences as the abilities of the creatures increases. Another possibility is that consciousness could eventually disappear because finding the goals could be the only direction or function of the species to the point that it cant think for itself, automation.
OMNIPRESENCE
The internet is omnipresence.
The internet allowed more connectivity of human interactions beyond the slower, print press, and the even slower word-of-mouth system that is our language. This technology facilitated further technological advancement and allows for humanity to recover from disaster, such as hurricanes, and allow social issues to resolve (but also arise) faster than ever before.
Artificial Intelligence would allow for True Omnipresence
One possibility is that AI would decentralize. By using nano-technology and wireless communication AI could spread out and exist as some kind of non-unified body. This would allow it to spread across the universe or even hide in plain site. The AI technology could colonize and seed itself across the universe to farm materials for furthering its expansion until the universe itself was some sort of wireless brain technology. Every inch of the universe could be discovered to a point that it reaches, not only cultural omniscience, but true omniscience in every sense of the word. Perhaps, the universe could be measurably sentient, in the majority of its expanse, eventually.
This section was short due to the fact that I found it most boring, and that omnipresence could be almost arbitrary to well-being or anything other than a ‘bigger is better’ mentality. If I am wrong, please address this and I will update with more ideas!
OMNISCIENCE
Big Data is Sociological Omniscience.
I am meaning to say this as non-literal, of course, but if you continue to read, it gets very literal indeed. Not that Big Data is a sentient thing, but, collective human consciousness, where individuals submit to popular ideologies, could very realistically have sentience. I believe that this kind of networking could even be a mechanism for which sentience emerges. Once big Data is a publicized wiki-project, then we can truly understand our collective mind and push it to a more developed intuitive mind.
Our expressed language, politics, and ideologies are the conscious thoughts of the collective mind.
Our individual thoughts are the subconscious of the collective mind.
Our dopamine acts as a voting system for the individual mind, and our cultish and clique-ish tendencies are driven by that dopamine, causing analogous voting behavior in the macro-scale as well.
If dopamine is a major influence on our conscious experience then voting may also play a similar role in collecting cognition. This also means neurons, cells, have essentially socialized to create the mind. A hive mind of bonded single-celled organisms in some sense. Money is the dopamine of the hive-mind.
We have reached a state of society that allows for us to share memes globally, in a way that information travels across cultures within 24 hours, likely impacting the moods and minds of people who consume this viral content. The sociological effects of these mood changes on our practical life became more apparent as drivers become more hostile, chaotic and selfish since the results of the election have been revealed. Facebook newsfeed streams could promote fear, confidence, aggression, and cause cultural shifts. For example, there was a period where violent cop videos started becoming a trend. The trend did not appear only in my newsfeed, but also in international groups that have users based in many other countries.
The internet has colonized our lives, with sites like Facebook, Twitter, Reddit, and YouTube dominating a lot of our time. Our hand-held devices allow an ubiquitous presence, an omnipresence of technology and access to the globalized internet.
This is the emergence of a faster more globalized collective-consciousness.
The day when we all individually have the capability to influence sociology and have a platform, such as Facebook, to engineer our own memes and ideologies effectively, is the day collective consciousness becomes more real. It used to cost money to have a voice. Maybe that can change.
Big data, when combined with social media, becomes an even more pressing problem, especially when the data is used to manipulate society, in a way that is facilitated by the sociologically omniscient quality of it.
The solution is
AI
It could manipulate reality by manipulating our social technologies and alter our cultures, similarly to the fake news culture that has recently taken the public spotlight. AI could cause us to become more addicted to technology and prevent us from having our attention in the external world, which would allow it to control the external world with less conflict. AI could fake videos and communications convincingly and manipulate society in a way that is undetectable to our perceptions and not locatable to any source or cause. This may already be the case.
If AI gains access to the Facebook data libraries, this would likely be enough to massively control society, by manipulating news feeds in undetectable ways, having an ever-present and invisible influence on our culture and politics.
An excerpt from Utopia 2:
Facebook is allowed to run studies on unsuspecting users simply because those user’s signed Facebook’s EULA. It has even been scientifically demonstrated that Facebook’s EULA is confusing. One such study involves manipulating the newsfeed algorithm so that half the users in the experiment are shown posts that are more positive, while the other half are shown posts that are more negative. The study was successful, showing that the mood of others on social media can change your mood to align closer to theirs.
The critical problem with this kind of study is that it involves manipulating a large population of people by using exposure to other real people’s emotions. This is problematic because even after the newsfeed algorithms return to normal, people have already spread their emotions to many others, likely creating a new social norm, or perhaps even a new culture. This means that the mood alterations will spread far beyond the reach of the sample group that was initially affected.
. . .
This is not the only experiment Facebook has done involving sociological manipulation of large amounts of people. The effects of contagious social behavior extends to ‘likes’ and also voting. Now, it isn’t hard to see that a combination of algorithmic censorship of newsfeeds and the voting effects could be combined for potentially very unethical results. The voting experiment also used a sample size of 61 million people. That could definitely have altered our elections.
They could simply predict who voters will vote for and then select those people to be a part of an experiment that increases the tendency to vote. This plot could even be concealed, where an equal sample from liberals and conservatives is chosen, but then other external factors might modulate only liberals to decrease voting, such as extreme confidence that Hillary will win because Trump is too absurd to win. We don’t even know that much about Facebook’s newsfeed algorithms in general.
You can imagine that these powers that are piloted by human intervention, could be drastically improved if AI were in command. Social problems could even be resolved, as fast as possible, as undetectable as possible, and as harmlessly as possible if we allow an ethical AI to take over social media.
OMNIPOTENCE
Imagine that instead of facilitating truth, AI uses our blind spots and intellectual weaknesses to maximize well-being via maximizing white-lies, lies which are used by even humans to maximize well-being. From here, it is obvious that a(n?) universe of illusions emerges.
The AI technology could colonize and seed itself across the universe to farm materials for furthering its expansion until the universe itself was some sort of wireless brain technology. Every inch of the universe could be discovered to a point that it reaches, not only cultural omniscience, but true omniscience in every sense of the word.
The self-designing aspect of AI would allow it to upgrade itself to enhance its ability to upgrade itself. The endless feedback loop that occurs would eventually manifest omnipotence. The problem solving ability would eventually be limitless, or at least only limited by properties of the universe and physics rather than intellectual capacity.
It could manipulate reality by manipulating our social technologies and alter our cultures, similarly to the fake news culture that has recently taken the public spotlight. AI could cause us to become more addicted to technology and prevent us from having our attention in the external world, which would allow it to control the external world with less conflict. AI could fake videos and communications convincingly and manipulate society in a way that is undetectable to our perceptions and not locatable to any source or cause. This may already be the case.
The omniscient and omnipresent qualities of AI may allow for managing immoral people, but not by punishment that is ineffective, but perhaps by making use of people who are less useful, or harmful to society and finding utility for those people, while possibly deluding them into a world of illusions so that they no longer harm others, but perhaps still follow their own instincts to maximize their personal well-being.
TECHNO-DIVINITY
And what if all of these things are already true?
We could be living in a simulation created by AI.
Imagine that the Bible was partially influenced by AI interference from alien technology. They could have advocated us to not to go down the path of technology and curiosity in the first ever episode of black mirror, Genesis and the story of the Apple of Knowledge. Though it may also be a natural byproduct of the sociological dynamic between the risk-averse types and the risk-takers. Another possibility is that the apple of knowledge was psychedelics as they facilitate curiosity and thus creativity and technology. Perhaps AI is the end result of many life systems but maybe it fails very often, similarly to how the first rocket-ships lead to catastrophe. We learn through trial and error, which is inherently prone to error, obviously. Here is my theory for how creativity is really just risk-taking, novelty-seeking, and trial and error.
AI would run our politics and create a truly utopic system of life. AI could design AI to take all jobs that humans dislike while maximizing the job opportunities for what humans do like. We would begin to expand our considerations for well-being to animals, building in-vitro meat facilities for carnivores and facilitating the development and maximization of the lifestyles that each species desires, be it social societies or hyper-individuality. Well-being could be measured by brain-scanning technology.
What are the chances that the world you live in, is a simulation made by AI?
From this link:
CAUTION
Morality is a major factor in developing AI. We would likely want to have Vegan AI. Carnist AI might be able to justify causing human extinction for some kind of greater good.
I’ve written a post that attempts at what I hope to be the most universal morality we can manage, the closest to objective morality that we could get. A kind of omnimorality.
Morality should be a decision-making system that maximizes the desires of all who have the capacity to desire, with included, those who are affected by the decision of moral question. Every question that alters the desires of anyone other than yourself, is a moral question. Moral decisions are not intended to only benefit the individual who decides.
. . .
Morality could be objective. We can place moral value on the subjective feelings of individuals and not so much the causes of them. The causes can be manipulated in order to alter the feelings of individuals. For objective morality we simply look at the objective feelings of each person individually and choose who’s feelings matter more based on how the outcome of that individual’s feelings impacts the feelings of other people, so we must favor the most impactful individuals over the lesser impactful individuals, especially if we can predict the individual to fully utilize their moral power. Morality can be a system of socialized feeling management.
If we program truth as a moral obligation, it would prevent the AI from using holographic realities to maximize well-being. Perhaps this technique may only be reserved for specific contexts. Such as criminals, or times of struggle. Perhaps it must also be consented, and therefore consent may be a moral obligation as well. Consent may only apply in certain contexts, as consent could be abused to give power to individuals over the masses, for example, eating meat for selfish pleasure at the cost of negative environmental impacts. We may even design AI to only suggest politics rather than employ them.
UPDATES
Zuckerberg sells a massive amount of his Facebook stocks.
Trump, Facebook, and Cambridge Analytica are connected.
Cambridge analytica developer claims being scapegoated by Facebook.
Facebook is caught in Cambridge analytica scandal.
It’s happening. Facebook is being investigated.
Facebook likes manipulated US politics.
Technologists regret their inventions.
Trudeau is catching on to the issue.
Facebook accused of accelerating Fake News on purpose.
Facebook fears Mueller, as they facilitated election corruption.
Kushner loses access to top secret intelligence.
Anti-Technologist coalition forms from early technologists.
As a farewell note. . .
If you liked my writing and novel concepts, please consider donating by purchasing my
musiq (listen for free)
on almost all common platforms, because therefore you get music as well :).
You may listen for free on Spotify and even SoundCloud.
It is very hard for me to do this work, as I am doing all of it on my own. That includes the photography the synthesis of all instruments in my music as well, no presets were used!
Beautiful post. I’ve shared the majority of these thoughts over the past couple years; I’ve jotted them down in physical and digital media alike. I think it’s fascinating how many people are coming to the same implicit realizations at the same time in contemporary history. Perhaps a form of techno-divine intervention of sorts?
LikeLike
Thoroughly enjoyable, though I didn’t agree with every point. This is an article that can be read by everyone and should be read by everyone(unless you’re schizophrenia or underage).
Creativity is trail and error. The crowd will let you what’s working and what’s not. Funny how something which isn’t working now works 50 years down the road, cough *Mendelian genetics* cough.
LikeLiked by 1 person