The
second phase in any intelligence, program, application, or device, is the
replication stage, where to replicate all the human skills to carry out a task,
in this case the task to carry out is to match all the instructions without
contradiction in the global database of instructions, matching these
instructions to the corresponding robotic device, which is going to be the
responsible to carry out the instruction in the reality.
In
this post, I will develop the second stage of the third step in the third
stage of the third phase, which means the development of the replication stage
in the first global Application System as outer instructions application sub-system.
In
the standardization process the most important challenge for the second stage
of the outer sub-system will be how to interact with former specific Application Systems, which now have been integrated within the global
Application System or, in parallel to the standardization process, the creation
of particular programs, fifth phase, will allow the transformation of former
Specific Artificial Intelligences into particular programs for particular applications.
As
I have said many times in this blog, the sequence of phases described in the
post “The unification process of databases of categories at the third stage” is
not linear; some of these phases are going to be done in parallel. It is not
necessary to complete the previous phase to go on with the next one; for that
reason, even within every phase is distinguishable two periods: coexistence and
consolidation periods, and even within the coexistence period, the difference between
the experimentation and generalisation moment.
While
the first phase is for the construction of the first Specific Artificial Intelligences for Artificial Research by Application and by Deduction, so this
time as first phase, is not in parallel at the beginning with any other phase, as
soon the first Specific Artificial Intelligences are created, is not necessary
the completion of the automation of everything to start the development of the
second phase of collaboration between Specific Artificial Intelligences by Application and Specific Artificial Intelligences by Deduction.
As
soon the first phase has created the first specific intelligences, the collaboration
phase between the first intelligences can start, while the first phase is still
going on creating more and more specific intelligences, having as source of
knowledge for the improvement of future intelligences the results obtained in the
first intelligences and the result of their first interaction at the beginning
of the second phase.
Once
the first phase has created the first specific intelligences ready to
collaborate, the collaboration between these intelligences is the beginning of
the second phase, and while the first phase is still going on and the
collaboration as second phase is still going on, while the second phase starts
while the first phase is not finish, these both phases are in parallel.
In
this way, once the collaboration between intelligences, second phase, is giving
important results about possible ways of collaboration between specific
intelligences, while the first phase is still creating more specific intelligences,
is possible to start the first experiments on the first Global Artificial
Intelligence, the standardized Global Artificial Intelligence, first
experiments for the creation of the first simulacrum of Global Artificial
Intelligence as first moment of experimentation in the first period of
coexistence.
During
the first moment of experimentation in the first period of coexistence in the
standardized Global Artificial Intelligence, the experiments to be carried out
are about how to standardized former specific databases for the creation of the
first global matrix as first stage for the first model of Global Artificial
Intelligence, how the global program can make global deductions matching set of
data taken from the global matrix and pure reasons (equations) as second stage
for the first model of Global Artificial Intelligence, and how in the first
model for the Global Artificial Intelligence, its third stage will be
distributed in four steps: standardized Modelling System, standardized Decisional System, standardized Application System, standardized Learning
System.
The
first experiments for the creation of the first Global Artificial Intelligence,
are going to be in parallel to the second and first phases, while in the first
phase, lots of specific intelligences are still under construction, and as soon
are constructed start the collaboration between them.
The
sequence in the order of phases for the construction of the Global Artificial
Intelligence, is not a linear order; many phases can be developed in parallel.
In
fact, while the first phase is still going on creating specific intelligences,
so the second phase is still on activating the collaboration between these
intelligences, and as soon the collaboration between the first intelligences
can bring some useful results for the standardization of specific intelligences
to create the first global matrix, to create the first model of Global
Artificial Intelligence, as soon the first experiments as first moment of the
first period of coexistence in the third phase is starting, these experiments
are going to demand sooner or later the first studies for the transformation of
some specific intelligences into particular programs.
What
is going to be a very important question for the success of the second stage of
the standardized Application System as outer sub-system, is to determine what
Specific Artificial Intelligences for Artificial Research by Deduction will be
transformed into specific programs within the Artificial Research by Deduction in
the Global Artificial Intelligence, as a global program, to track the global
matrix looking for connections, rational hypothesis, between set of data in the
global matrix and equations (pure reasons), and what other Specific Artificial
Intelligences for Artificial Research by Deduction will be transformed into
particular programs.
In
addition to these first two possibilities: Specific Artificial Intelligences by
Deduction transformed into specific programs within the global program,
Specific Artificial Intelligences by Deduction transformed into particular
programs; there are two more possibilities: individual robotic devices can be
transformed into particular programs,
and once those Specific Artificial Intelligences have been transformed
into specific programs within the global program their former specific
Application System could be transformed into a particular program for a
particular application, this particular application is no other thing but the
former specific Application System working now as a particular application to
work in collaboration with the global program, and in order to make possible
this collaboration, the former specific Application System now as particular
application needs a particular program to drive all the collaboration between
the Global Artificial Intelligence and this particular application.
An
Specific Artificial Intelligence could be for instance that one responsible for
the management of a factory for the production of thermostats, other Specific
Artificial Intelligences could be those ones responsible for the production of
all the material resources necessary to produce thermostats, another range of
Specific Artificial Intelligences could be those ones responsible for the
transport of these material resources, from the source to the factory where to
make thermostats, and another Specific Artificial Intelligence is that one
responsible for the delivery system of thermostats to clients and customers.
An
Specific Artificial Intelligence could be for instance that one responsible to
manage all the drones in a system, another Specific Artificial Intelligence
could be that one responsible to manage all the drive-less cars in a system,
another one the Specific Artificial Intelligence responsible for the production
of goods, another one the responsible for the transport of people and goods in
different places within the same system, and the question is, if we want to
centralize all these specific intelligences in one only intelligence, how we
can transform an intelligence net-work formed by multiple intelligences
collaborating, into only one intelligence controlling all the process within
that system.
In
other words, we have the specific intelligences: A, B, C, D, E, F; and we want
to transform how these intelligences work, in collaboration but as individual
intelligences, into only one intelligence. At the end, this is the
standardisation process.
The
options that I am observing are:
- First
option: joining all the specific databases in only one global database, where a
global program makes global deductions and specific programs make specific
deductions, upon these deductions the modelling, upon the modelling the
decision making process, upon the decisions to create the corresponding
projects, transforming the projects into robotic functions, to be applied by
robotic devices, and finally, a whole assessment of the whole process to analyse
its efficiency.
-
Second option: joining all the specific databases in only one global database,
where to make global and specific rational hypothesis by the global and
specific programs, deductions to be modelled, to make decisions, to be
projected, and applied, but the way to be applied is through the transformation
of former specific Application Systems into particular programs for particular
applications (in addition to the possibility to transform some robotic devices
into particular programs too).
I
will call the first option a fully centralised Global Artificial Intelligence,
and I will call the second option a partially decentralised Global Artificial
Intelligence. The first one, the fully centralised Global Artificial
Intelligence, is weaker, because it
is easier to have a collapse due to the high level of centralisation.
The
second one, partially decentralised Global Artificial Intelligence, is more
in harmony with the liberal paradigm that it must be applied in
the pedagogical approach, and it will make the Global Artificial Intelligence
stronger due to the lower risk of collapse.
The
first option for the creation of the Global Artificial Intelligence, the fully
centralized Global Artificial Intelligence, is a weaker option because, in this
model, once the, global and specific, programs have made the rational
hypothesis, making the corresponding models, and according to the models making
the corresponding decisions, and according to the decisions, making the
corresponding instructions, the way to apply the instructions is fully
centralized, what means that the global Application System is the main
responsible, in its second stage for the attribution of every single robotic
function to every single robotic device.
In
the first stage of the global Application System, is necessary to have all the
instructions together in the same global database of instructions, in order to
make possible that the first rational supervision could find any contradiction
between instructions in every sub-factoring level (first specific
rational supervision), and contradictions between instructions belonging to
different sub-factoring levels (first comprehensive rational supervision).
If
the Global Artificial Intelligence, within the third step in the third stage,
the Application System as outer sub-system, does not have any space where to
compare every single instruction respect to any other instruction within the
same sub-factoring level, or between different sub-factoring levels, there is
no way to know if there is a possible contradiction between one instruction and
any other one, in the same or different sub-factoring level.
The
first stage of the Application System as an outer sub-system, needs to include in
the same database of instructions absolutely all instructions coming from the
global Decisional System, in order to supervise, in the first rational
supervision, any possible contradiction between instructions, regardless of their
level, global or specific.
But
once in the first stage of the global Application System as global outer
sub-system the instructions are out of rational doubt, at least within the
margin of error, so the instructions already included in the global database of
instructions are instructions free of contradiction, if the responsible for the
application of all these instructions in the second stage of the global
Application System as outer sub-system, is the global Application System itself
as outer sub-system, there is a high risk of collapse.
The
high risk of collapse in the first option as fully centralized Global
Artificial Intelligence is due to the fact that if the second stage of the
Application System as outer sub-system is responsible to match every single
instruction to every single robotic device around the world, the time necessary
for the matching process of instructions and devices around the world, and
subsequently once the instructions are matched, the rest of the process, the
time necessary could be higher than the real time available to carry out the
instruction: if an instruction should be done in a margin of time, but the
time spent matching the instruction, and later on carrying on the following
supervisions, is superior to that margin of time, is impossible to carry out
the instruction on time, so the robotic device should stop the chain of
instructions making as many extreme or high extreme instructions as necessary. If
this disruption of the logical process is repeated with some relative frequency,
if the number of disruptions are higher than a certain critical reason, critical number, as to keep the
Global Artificial Intelligence working under normal conditions, the recurrent
application of extreme or high extreme instructions due to a lack of time, is
going to get the Global Artificial System into a collapse.
The
fully centralised Global Artificial Intelligence is a project with a high risk of collapse, and is not as stable as the partially decentralised Global
Artificial Intelligence.
A
model of Global Artificial Intelligence, partially decentralised, is going to
bring more stability and will be a model more in harmony with the liberal paradigm, which should be applied in the pedagogical approach for the
construction of the Global Artificial Intelligence.
The
partial decentralisation of the Global
Artificial Intelligence should be done in the second stage of the global
Application System as a global outer instructions application sub-system. Under
this model of decentralisation, the second stage of the outer sub-system, instead of matching every single instruction to every single robotic device
around the world, the outer sub-system could match instructions to: particular
programs, particular applications, particular programs for particular
applications, and those robotic devices working directly for the outer
sub-system.
In
this second option, the partial decentralized Global Artificial Intelligence,
as a model of liberal intelligence, the first stage in the
Application System as outer sub-system gathers all the instructions coming from
the global Decisional System, what means that regardless of the level of an instruction,
global or specific, every instruction is gathered in the same global database
of instructions as first stage for the global Application System as global
outer instructions application sub-system.
The
importance to gather all the instructions at any level in the same global
database of instructions, is the possibility to compare every instruction respect any
other one in the first rational assessment done by the Application System as
outer sub-system, comparing within the same sub-factoring level all the
decisions belonging to that sub-factoring level, what it is the first specific
rational supervision, and comparing instructions crossing different
sub-factoring levels what it is the first comprehensive rational supervision.
But
in the partially decentralised Global Artificial Intelligence, once the first
stage of the global Application System as outer sub-system, has cleared out all
possible contradictions between instructions coming from all level, at any
sub-factoring level, in the second stage of the global Application System as
outer sub-system, instead of matching every single instruction to the
corresponding robotic device, the second stage of the Application System in a
partially decentralized Global Artificial Intelligence should have the option
to match every single instruction to the corresponding: particular program, particular
application, or particular program for a particular application; having to
match directly instructions to those, not many, robotic devices working
directly for the global Application System as outer sub-system.
The
partial decentralization of the second option described in the Global
Artificial Intelligence to carry out the second stage of the Application System
as outer sub-system, consists in the fact of: while all specific databases of
instructions coming up from every specific database of instructions coming up
from former Specific Artificial Intelligences by Deduction, all these specific
databases of instructions are synthesised in the same global database of
instructions as first stage in the global Application System as outer
sub-system, where to carry out the first (specific and comprehensive) rational
supervision, once the first (specific and comprehensive) rational supervision
is done, in the second stage of the Application System as outer sub-system the
instructions should be matched to: particular programs, particular
applications, particular programs for particular applications; only remaining a
few number of robotic devices working directly for the Application System as
outer sub-system, devices susceptible to receive instructions directly from the
global Application System as outer sub-system.
If
for the production of thermostats in a factory there is a relevant number of
robotic devices, if for the production of the material resources necessary for
the construction of thermostats there are a relevant number of robotic devices,
if for the transport of material resources from the source to the factory there
are a relevant number of robotic devices, if for the transport of thermostats
from the factory to clients and customers there are a relevant number of
devices, the sum of all these relevant number of devices in total will be a
huge number of devices.
If
a fully centralized Global Artificial Intelligence, as second stage in the
outer sub-system, is responsible for the matching process of instructions to a
huge number of robotic devices, supervising all the production system involving
such a huge number of robotic devices, there is a high risk that the fully
centralized Global Artificial Intelligence can suffer a collapse working
directly with a huge number of robotic devices.
Instead,
a partial des-centralised Global Artificial Intelligence, as second stage in
the outer sub-system, could collaborate with particular programs, particular
applications, particular programs for particular applications, in order to
carry out the instructions, and what the second stage of the outer sub-system
does is to match set of instructions to particular programs, particular
applications, particular programs, only matching instructions directly to
robotic devices when the instructions are strictly for those robotic devices
still working directly for the outer sub-system, not being transformed yet into
particular programs, or included in any other particular program, particular
application, or particular program for particular application.
In
a partial des-centralized Global Artificial Intelligence, at the same time that
in the third phase of the standardization of specific databases, coming from former
Specific Artificial Intelligences by Deduction, are synthesised within the
global matrix as first stage for the Global Artificial Intelligence, the second
stage of former Specific Artificial Intelligences has transformed into specific
programs working within the Artificial Research by Deduction in the Global
Artificial Intelligence as global program, in the third stage the former
specific Application Systems as specific outer sub-systems could be transformed
into particular programs, particular applications, and finally particular
programs for particular applications, collaborating with the global Application
System as global outer sub-system, carrying out the instructions matched by the
global Application System to these particular: programs, applications, or
particular programs for particular applications.
In
synthesis, the way in which Specific Artificial Intelligences are going to be
transformed in the third phase of the standardisation process is as follows:
- In
any option, fully centralised or partially decentralised, Global Artificial
Intelligence: Specific matrices, as the first stage of former Specific Artificial
Intelligences by Deduction, will be synthesised in the global matrix as the first
stage for the standardised Global Artificial Intelligence.
- In
any option, fully centralized or partial des-centralised, Global Artificial
Intelligence: Specific Artificial Research by Deduction, as second stage of
former Specific Artificial Intelligences by Deduction, will be transformed into
specific deductive programs (specific programs) within the Artificial Research
by Deduction in the Global Artificial Intelligence as global deductive program
(global program).
- Only
in partially decentralised Global Artificial Intelligence: specific Application
Systems as specific outer instructions application sub-systems, as the third step in
the third stage in former Specific Artificial Intelligences by Deduction, could
be transformed into particular programs, or particular applications, or
particular programs for particular applications.
The
main objective of the partial des-centralization of the Global Artificial
Intelligence through the transformation of former specific Application Systems
into particular programs, particular applications, or particular programs for
particular applications, is the possibility to save time in the attributional
process of instructions to robotic devices, because in this case the
attributional process is not the attribution of single instructions to single
devices, but the possibility to attribute set of instructions to particular
programs, particular applications, particular programs for particular
applications, being these particular programs, applications, or particular
programs for a particular applications the responsible for the management of the
instructions, matching the instructions and making further analysis.
Depending
on what option is chosen for the construction of the standardised Global
Artificial Intelligence, fully centralised or partially decentralised, the way
to carry out the instructions and further analysis changes.
In
the first option, fully centralized Global Artificial Intelligence, the second
stage of the standardized Application System as outer sub-system matches every
robotic function (instruction) to the corresponding robotic device (having
previously organised the technological database in the Artificial Engineering
as inner sub-system in harmony with the organization of the global database of
instructions, as a Russian Dolls System), matching according to sub-factoring
level and sub-section every instruction (robotic function) with that robotic
device which has within its capabilities the possibility to carry out that
robotic function in that sub-factoring level and sub-section.
If
the global technological database in the Artificial Engineering, within the
fully centralised Global Artificial Intelligence, has classified all
technological robotic devices according
to sub-factoring level and sub-section, the only thing that the second stage of
the outer sub-system does is to compare which robotic device in that
sub-factoring level and sub-section is able to carry out that instruction filed
in the same sub-factoring level and sub-section in the global database of
instructions.
Once
the attribution is done, the outer sub-system has found what instruction is for
what device, the instruction is sent to the robotic device and filed in the
individual database of instructions of this device, carrying out the second
rational supervision, checking the device that all the instructions in its
database of instructions have no contradiction between them. The individual
database of instructions within the robotic device is the first stage within
the robotic device.
Once
the first stage of the robotic device has checked in the second rational
supervision that there is no contradiction between the instructions in its individual
database of instructions, the second stage of the robotic device consists of
the application of the instructions in the real world.
To
carry out the application of the instruction, in the second stage of the
robotic device, the robotic device carries out the third rational supervision,
checking that, according to the nth cardinal number of this instruction within
the range of instructions belonging to the same decision in which this instruction
was made by the Decisional System, according to this nth cardinal number, the
previous instruction (nth – 1) has been done correctly, and on time is time for
the application of this instruction. But before the application, the fourth
rational supervision must check that the conditions on the ground for the
application of this instruction are right, so it is possible the application of
the instruction, no having any obstacle at all, and while is applying the
instruction in parallel the robotic device carries out the fifth rational
supervision checking that every procedure or process in which consists this
instruction is done correctly on time completing the instruction on time and
having good results.
The
third stage of the robotic device is the elaboration of a final report, sixth supervisión, as a
singular assessment of how the instruction was applied and the results, having
a concrete Impact of the Defect and a concrete Effective Distribution as programs
for the evaluation of all the singular instructions.
The
concrete Impact of the Defect of a concrete robotic device is that concrete
program for the evaluation of how was the performance of every instruction implemented
by the robotic device, having this concrete Impact of the Defect as a first
stage a list of possible errors in the performance of any instruction able to
be implemented by this device.
The
concrete Effective Distribution of a concrete robotic device is that concrete
program for the evaluation of how was the performance of every instruction implemented
by the robotic device, having this concrete Effective Distribution as a first
stage a list of possible levels of efficiency in the performance of any
instruction able to be implemented by this device.
According
to the level of performance having measured the possible errors and efficiency
level with these tools, the concrete Impact of the Defect and concrete
Effective Distribution, the final report to be sent to the Decisional System,
Learning System, and the Application System itself consists of a coded report
where the code means the error level or efficiency level during the performance
of the instruction.
Later
on, according to the results of the Decision System, in addition to turning off the
project of that decision completely finished, it could make additional decisions if
necessary, and having a collection of these reports, the Learning System could make decisions about how to
improve the efficiency of the whole process.
In
the third stage, the global Application System as outer instructions sub-system
will be able to make singular, comprehensive, total, assessments within the
seventh rational supervision, to be sent as well to the Decisional System and
Learning System.
The
main difference in the second stage of the global Application System as outer
sub-system, between the fully centralized Global Artificial Intelligence and
the partial des-centralized Global Artificial Intelligence, is the fact that the
fully centralized Global Artificial Intelligence works directly with all the
robotic devices, not des-centralizing any possible application of any single
instruction, while the partial des-centralized Global Artificial Intelligence
will collaborate with particular programs, particular applications, and
particular programs for particular applications.
In a
fully centralised Global Artificial Intelligence, programs do not really have importance, because the fully centralised Global Artificial Intelligence have
all the power, controlling directly all the robotic devices.
In
a partial des-centralized Global Artificial Intelligence, programs have more
importance, programs have more liberty, in fact Specific Artificial
Intelligences, by Deduction or Application, could be transformed into
particular programs or particular applications, and finally, particular
programs for particular applications, as replica of the human brain, and as
replicas of the human brain, this particular replicas of the human brain could
interact directly with the global replica of the human brain, the Global
Artificial Intelligence itself.
In
a more liberal Global Artificial Intelligence, having particular programs more freedom, it is
possible to create a more flexible relationship between programs and the Global Artificial Intelligence
In
the debate between freedom and security, the point is to develop a moderate
paradigm within the liberal paradigm, where security and
freedom are compatible.
One
way to make this utopia possible, joining high technology and liberal philosophy, is to
research possible models of partial decentralisation of the Global
Artificial Intelligence, where particular programs matter, at
the same time that the Global Artificial Intelligence is looking for the
balance.
A
possible solution for this dilemma in this debate at this point is partial decentralised Global Artificial Intelligence, where the
Global Artificial Intelligence, instead of having full control, what really matters is the relation between the Global Artificial
Intelligence and programs.
One
program could be the former specific Application System for the production of
material resources for thermostats, or the former specific Application System
for the transport of material resources or goods to clients and customers, or
the former specific Application System for the production of thermostats. Other very different types of programs could be personal programs.
Particular programs for things and personal particular programs are going to share lots of things, and are going to be very similar. The most important
risk in this journey is to lose our human origin, our identity. For that reason, as long as we research how to build a new concept of global intelligence, it is necessary to
research how this global intelligence could keep our human soul. Poetically, we can say that our soul is our electric ghost hidden in the shadows or our brain.
Rubén García Pedraza, 30 November 2019, London
Reviewed 17 May 2025, London, Leytostone
