My point of view on AI

From creatures to creators

It has been a while since I write about AI. From promise to delivery, we are now entering the stage of full adoption - where we are passed the amazement of this new fire given by the spirit of deep learning - into a full battle mode where implementation is the order of the day.

At the same time, it’s already the common knowledge the profound impact of AI as a deeply transforming technology that not only willl it deserve an index number in the charts of the Industrial revolutions (5th in this case) but also as a pivotal moment in the chronology of human history.

Many portrait this pivotal given AI’s capabilities to create itself. A self propagating machine, that not only creates things, but it also creates itself - replication and iteration. This is called a self-reinforcing loops, or more commonly positive feedback loop, where an initial action triggers a response that amplifies the original action.

[Self reinforcing loops diagram]

We have many examples of this in our world, although they are so

• Biological Evolution, Scientific Discovery, Software Development, Capital Accumulation, Social Media Virility,Neural Plasticity… all these action/process enhance their process itself.. building a cascade effect, that will have an exponential shape.

All in all, we tend to wrap up this phenomenon in one catch phrase “winner takes it all”. So, can we already foretell AI as the big winner? Yes and No. We would have to go a bit quantum in order to describe the winning and loosing concepts of AI. Two states at the same, where in one AI is winning and another where it’s not. And they can coexist in a superposition, but as the observer we can only see one outcome. It’s important then to reason on what “winning” means - and here is where AI might have it’s loop hole.

[Quantum mechanics graph of superposition of states in AI]

Modern day "Winning" vs "Loosing"

So what do we mean by “winning“ and “loosing”. Both can be very broad in what they encapsulate - but on the first order of perception - I would say that everyone is worries about an AI that will surpass and control humans - in an unequal manner, where no human can even compete to overcome its intelligence, power and capabilities. So, we can summarise an AI winning as “total dominion over human species and full autonomy to thrive”.

On the other hand, a loosing AI it’s one that is subordinate to humans, where its objective is solely bridge human connection and fulfil tasks on demand, always within the limitations of the human mind.

I believe the future will be a superposition of states, where both situations will coexist - on one hand humans will drink from the wonders of AI, and on the other hand - AI will thrive on it’s own into domain far from the natural course of human conscience.

It is when we force an observability state that these worlds will connect and be enforced into one state.

The cataclism view and the complexity of goals

The doomsday advocates typically give arguments where this intelligent being might see humans as an obstacle for some higher ground obstacle that it might have. And hence, this system might kill all humans to achieve that. That is legitimate hypothesis, but I would put is a simple one. That assumes that this super intelligent system that is being developed is grounded on a single prompt with a binary success criteria - which given the complexity that it has, wouldn’t make any sense. As the complexity of an intelligent system evolves, so does the complexity of its goals. From the real of the purely quantitative we enter into the qualitative realm - in which humans have been existing for a long time. And as such, I don’t see it as a super plausible outcome - the extermination of humans.

But this arises a much deeper question, which is the underling value that will be given to humans in general and until what level will they be replaced by machines. Looking at Teillard de Chardin’s elemtan movement of life -

• Reproduction: form one making two
• Multiplication: growing communities, building something bigger than ourselves
• Renovation: regeneration of our own being (physically and mentally)
• Conjugation: variabilty and adaptability to survive
• Association: Creation of ideas a higher complexity
• Controlled Activity: Adapt activity to the outer world, como One can start drafting the excess load that humans have been putting on top of these very very basic biological needs, and we can restart outlining a new vision for humanity. In the end, all we want is to reproduce and multiply in order to create something bigger than us. We do this via conjugation and association, adapting to the environment in which we exist through some control mechanism of our own activity. All this, can still and will be done in a paradigm where AI is embedded In all stages of our journey - just like electricity 100 years ago was engrained in our lives.

But we haven’t answered the question - how far will AI replace humans? What will be left? It’s very hard to give a straight answer, but we can start answering this question by looking at some examples. Look at chess. In 1997, Deep blue - a computer designed by IBM defeated Garry Kasparov. An impressive feat, and certainly the end of the AI winter. However, this did not put an end to our interest in looking at humans playing chess. On the contrary - it only gave emphasis to that human limitation, after knowing that there algorithmic methods to outmatch a chess player at any level. We filtered out the mathematical/probablistic curiosity of the game, into the exploration of human capacity. In a way, we strive for the overcome of our own limitation when watching sports - and that limitation is necessary and needs to be of the same nature as us. Now, on the statues of this limitation, we can elevate through the ranks of the various limitations into a higher order limitation, more on the spiritual level. And hence advocate that the a man on a pst AI age, will strive for higher grounds of spirituality and conscious as the ultimate layer of limitation exploration. Her will be where compassion, empathy and understanding will be manifested - as these are the realms that are not (yet) being coded into ai - since they do not exist in our written corpus - only recounts of their experiences.

So, in summary - our goals will be shifted to momentanoues task oriented goals, into higher realms of spirituality and understanding. From created to creator, humans are now observers of a self driving creation.

Fake jobs and displaced societies

What we do not look for - is the existence of what we do not consider a limitation. Or at least what a substantial part of the population does not consider as a limitation. This is the concept of the so called “fake jobs” - where jobs apparently simple and not customise to ones character are more of a social placement, or time fillers for individuals whose true impact is more on a social level than on the practical level. Some people’s roles in the great scheme of societal florishment have more of social roles than active ones. These people are great contributors by the act of being friends, parents, or just audience for the thrive of ideas. One of the key aspects of the human phenomena is its social aspect, where we see in each other images of ourselves and we build ideas on ourselves and our community by observing each other. As such, not the impact of one into the society can be very big just by the fact that one exists.

Now, will AI overthrow these fake jobs? Filling out data into forms, call centres, personal assistants, project managers, etc etc - in essence - jobs with little intelectual context windows and conceptually easy to automate. Absolutely - but not everywhere. These jobs have a very direct function of fostering organisations by having people attached to it - and creating a mass of employees whose function goes beyond merely productivity. The larger a company is, having multiple employees is a way of governance and brand trust that gives compnaies edge in the market and keeps idle time within reason. Also, the sense of competition is also importante for companies to achieve innovation - but above all, the Pareto distribution o f 80/20 will have a greater 20% the bigger is the workforce.

Despite all this, certainly companies will strive for workforce reduction - but only temporarily until new fake jobs arise. These fake jobs will be done within a higher layer of abstraction, already embedded with AI and within a paradigm of human abstraction within an AI first world.