PICHUTZ: THE SHAPE OF KNOWLEDGE
Vo Viet Anh, XPRIZE Vanguard, firstname.lastname@example.org
This document provides a high-level description of Pichutz (formerly “Singularity Pyramid”), an
online multidimensional knowledge mapping platform (which, for simplicity, can be regarded as
a futuristic Wikipedia), wherein each unit of knowledge is associated with a probability
distribution of probabilities to reflect the probabilistic nature of reality. Knowledge mapping
wouldn't be a big deal if not for four well-defined abstract dimensions (Conceptual, Purposeful,
Scopic, and “Language”) that capture the collective body of knowledge in its entirety. With the
Singularity placed at one end of each dimension as the ultimate goal, these dimensions give you
a sense of direction in learning/creating knowledge.
With knowledge mapping, Pichutz provides Singularitarians with a bigger picture to realize the
roadmap towards the Singularity while seeking to address numerous problems that are impeding
their progress—no coherent knowledge tree structure for academic research and massive
collaboration, no strategic profiling platform, leader discord, lack of awareness of existential
risks, negligence of macro-vulnerabilities, “language” barrier to pure knowledge and redundancy
of knowledge. It is also a tool that I myself have always coveted for use in planning, decision
making, risk management, and resource allocation. The “theory of weaknesses” based on Pichutz
has the potential to change your worldview forever.
Engineering Pichutz is an unprecedented and formidable task which I foresee rapid advances in
technology in the next year may fulfill. Also, the intuitive (easy to memorize, visualize, and
contribute) and functional (designed for practical use) structure of Pichutz makes it a candidate
for the brain of the Singularity Superintelligence (SS). Note: there is no technical detail in this
Keywords Singularity, knowledge mapping, hierarchy, probability, crowdsourcing
The model of knowledge for Pichutz (pronounced pē-shūt'z) featuring 4 dimensions of
knowledge is the result of my lifelong attempts to make sense of all knowledge I had stumbled
upon. It is the ultimate culmination of all of my intellectual and emotional experiences to date,
the answer to all of my questions
, and the tool to solve every future problem.
4 dimensions of Pichutz were defined for a singular purpose of reaching the Singularity:
(1) The 1st
(Conceptual) dimension shows the conceptual hierarchy of knowledge. It places
the Singularity tentatively at the top as the goal to climb towards
(2) The 2nd
(Purposeful) dimension draws the lines between 3 classes of knowledge based on
their relevance to our ultimate goal of survival. It forces you to ask yourself “What good
is knowing this for?” and reminds you to maximize resource efficiency by focusing on
knowledge at the Core
(3) The 3rd
(Scopic) dimension groups up knowledge. It encourages you to combine branches
of knowledge into more substantial knowledge
such as sophisticated technologies, and
compel you to back your knowledge with data and facts.
(4) The 4th
(Language) dimension presents different versions of the same knowledge. It
welcomes you to provide and translate knowledge into rawer, less ambiguous, more
machine-friendly forms of language (other than human language).
After mastering these dimensions, you will organize your thoughts and look at the world in
entirely different ways
. With them, Pichutz seeks to address 6 mains problems:
1. Academic research and massive collaboration: Web users can only look at one piece of
knowledge at a time (at Wikipedia or other Google search results) rather than a big picture
containing all pieces of the collective human knowledge and their relations. I grew dissatisfied
with what I could find during online research with Wikipedia
. It was not because
there was too much knowledge on the web but because a single coherent tree structure for all
knowledge was nonexistent. It was easy to get lost trying to figure out all foundations of a
complex multidisciplinary knowledge (e.g. cloud robotics) or all applications of a theory (e.g.
general relativity), the highest level philosophical insights of experts in a field (e.g. “Bitcoin will
be bigger than the internet”) or the deepest level technical tips (e.g. algorithms measuring
Levenshtein distance between two strings), or in which intellectual direction I should explore in
order to ultimately invent new knowledge, i.e. original research or novel technology. At the same
time, scientists lack an optimal tool that integrates the tree structure of knowledge for massive
collaboration in solving hard problems1
. To address these problems, I envisioned a futuristic
Wikipedia that surpasses the conventional concepts of all knowledge bases
, make it easier to
research systematically, navigate intellectually, collaboratively solve hard problems to arrive at
original and/or profitable ideas.
2. Strategic profiling: There are too many credentialed people out there to decide who to listen to,
partner with, or hire. Every person is represented by a long list of achievements, companies,
publications, public profiles, questions and answers, social network, status messages scattered
across multiple sources on the web, which take too long to research. On the other hand, what we
are concerned the most with when gauging someone’s skill depth, wisdom, and mindset/logic,
are (1) his deepest technical knowledge, (2) highest level conclusions, and (3) how one jumped
from assumptions to conclusions (in an extreme case, Deepak Chopra’s logic was so flawed he
made long shots to inane conclusions outside of his fields). By treating each user as an object on
dimension (together with other group objects) comprised of pointers to all pieces of
knowledge he has created, endorsed, learned, and connections he has drawn between these pieces,
Pichutz allows each user to maintain a singular profile on the web including the most
meaningful information about that person’s mind—basically what matters most to find or be
found by like-minded partners, employers, or benefactors.
3. Leader discord: It’s very common to find experts in the same field with glaringly contrasting
and there’s no way to verify their conclusions in order to decide which one is
. For example, Dr. Kurzweil complained in [Kurzweil 2012] about Paul Allen’s lack of
point-to-point critique that preluded any constructive argument. Each of them stood for a unique
set of data, facts, life experiences, influences, and disciplinary knowledge that simply doesn’t
overlap with the other’s enough to convince the other. There’s no platform for them to map their
“knowledge set” to serve as the basis of their argument. In other cases, I often encountered
intellectual leaders who were apparently smart with deep expertise in their respective fields but
somehow naïve or even ignorant regarding worldly matters
. Some were too clouded by their
ego (inability to admit one's own fault lest losing face) or emotion (traumatically forged reaction)
to reason beyond their jelled assumptions, to unlearn and relearn. A constant question in my
mind was “What exactly made this person assume that? Does he know a crucial piece of
knowledge that may change my mind, or simply have a ridiculous Deepak-Chopra-esque logic?
Where is he stuck?” Without such knowledge, there’s no way to pinpoint the root of their
assumptions and hold a constructive conversation. As a result, I had been thinking of a platform
to help mitigate bias and close the gap between subjectivity and objectivity
can exhaustively trace their current positions back to assumptions that constitute their
fundamental differences, and verify these assumptions in order to reach a consensus.
4. Unawareness of danger: A level of awareness high enough to perceive existential risks
requires a high amount of the right knowledge. Since any knowledge has a price-tag, there is just
not enough resource to pursue all kinds of knowledge. However, only a limited number of people
at this level of awareness (e.g. those at CSER3
) are concentrating their resources on
acquiring the core body of scientific and technological knowledge that may help us avert these
Centre For The Study of Existential Risk http://cser.org/about/who-we-are/
The Future of Life Institute http://thefutureoflife.org/who/
risks, while the majority of the world (even smart people5
) is ignorantly6
(including their brainpower7
) on knowledge that doesn’t directly translate to our long-term
survival. A series of frustration had led me to the conception of Pichutz: When I became
conscious of climate change and other environmental issues, my peers in university hadn’t, so I
joined the environmental movement alone. When I became conscious of the limitations of purely
environmental solutions and approaches, my fellow activists and scientists hadn’t, so I engaged
in the Singularitarian community to explore knowledge in various fields alone. When I realized
that the community, in their fixation on disruptive technologies, had lost track of the big picture
where politics and public ignorance largely affect the pace of their progress towards the
Singularity, Singularitarians didn’t seem to heed, so I conceived Pichutz.
5. Negligence of weaknesses: Everyone in the computer sciences knows how much damage bugs
or deep technical vulnerabilities may cause to rapidly expanding information systems. However,
very few are aware of how much threat larger-scale and more abstract vulnerabilities pose to our
survival against the odds of existential risks from Nature or even our own exponential
technological revolution. To ensure our survival, we must learn to identify vulnerabilities (called
“weaknesses” in this paper) from the smallest to the largest scale—every bit and piece of
information we miss, the content of what you eat or breathe, any possibility we fail to take into
account in our plan/theory/algorithm, or any failure to acknowledge and insulate our own
biological limitations, and the very population of people that carry such weaknesses. With a
knowledge map, it would be possible to circle where weaknesses might be most damaging if
exploited in order to contain them with layers of high-level security plans, or concentrate
resources to eliminate them once and for all—a process called weakness management. Against
the prospect of future crime, Pichutz is meant to be worth millions of words on security8
6. Language barrier and knowledge redundancy: Wikipedia is written in the most human-friendly
language—human language, which is not optimal for knowledge representation. IBM’s Watson
and most AIs have to derive knowledge through crawling and parsing countless data in this
language, which might take indefinitely long for them to achieve human-level intelligence. So,
why is there no Wikipedia for AIs, written in more machine-friendly languages? Humans expect
machines to understand them and yet so few (mostly programmers) can make their thoughts
clear-cut enough to be understood by machines in forms of code and algorithms. There are code
crowdsourcing websites such as Stack Overflow, TopCoder, and Kaggle but their solutions are
so disconnected and limited to very deep technical problems, while there is an enormous amount
of high-level knowledge that can be translated into more machine-friendly languages by just
anyone (not just programmers) so that machines can help us solve higher level problems.
Another problem with our reliance on human languages, which are a barrier to pure knowledge,
Having a bunch of best thinkers become lawyers is waste of human capital http://qr.ae/qW6rp
is the explosive redundancy of knowledge, e.g. “bin Laden might be dead” and “bin Laden was
killed.” If there is a platform to crowdsource systematic codification of hierarchical knowledge,
we can accelerate machine learning (a top-down approach to complement current bottom-up
approaches), all while downsizing the collective knowledge, and teaching users algorithmic
thinking to think in first order logic and rigorously define actionable high-level knowledge, e.g.
an algorithm to allocate resources and determine actions taken against specific populations
within a time frame that guarantees existential risks lower than a threshold. The 4th
a future project to upgrade Pichutz into such a platform.
Figure 1: Conceptual illustration of Pichutz
At first, this model had many names—the Pyramid of Knowledge, Symmetry, Wisdom, etc. In
2012, I learned about the Technological Singularity and realized that this model could become a
powerful tool to achieve it faster, if not in time. Hence the name Singularity Pyramid to align my
goal with the Singularitarian community’s. After a serious naming effort, it’s Pichutz now.
I tried to make Pichutz easier to understand by referencing Wikipedia to imply that this is also a
"knowledge system." However, this platform is not an encyclopedia. Definition of hot dog,
birthplace of Barack Obama, or details of a music album, is among what I LEAST expect can
make it into Pichutz. Because with Pichutz, I'm not just proposing a new knowledge tree
structure, but an entirely new way of filling one. Its knowledge is need-based, meaning
knowledge is filled in to either create new knowledge or prove/disprove the existing, NOT to
collect all existing knowledge in one place.
For light-hearted examples, “seawater is abundant on Earth” + [seawater-based fusion energy
generation technologies] [laws in a seawater economy] (new knowledge); data, facts, life
experiences, and disciplinary knowledge are posted to (dis)prove "AI could spell the end of
mankind" by Stephen Hawking or "Only morons start a business on loan" by Mark Cuban.
Compared to Wikipedia, Pichutz also takes time to crowdsource knowledge, but it’s more
functional, offering a wide range of applications in everyday, profit-driven, or high-level
, and thus promises benefits for contributors and ROI
Table of Content
II. The rooms 7
III. The 1st
IV. The 2nd
V. The 3rd
VI. Secondary dimensions 15
VII. Terminology and abbreviation 17
VIII. Building the pyramid 18
IX. Application 25
X. Theory of weaknesses 27
XI. My next step 30
XII. Interpretation and notes 31
XIII. Example 38
XIV. Bibliography 51
XV. Vitae 52
XVI. Dark 52
II. THE ROOMS
A unit of knowledge that can be (NOT “meant to be”) mapped on Pichutz is not limited to
academic knowledge but refers to the broadest possible definition of knowledge—anything that
can be known, which is as infinite as the universe itself. It can be anything conceivable: a qubit, a
datum, a fact, an opinion, a citation, an object, an idea, an action, a thought, an event, a
phenomenon, a theory, a plan, or virtually any information to predict exactly anything that may
happen the next moment, etc.
Metaphorically speaking, Pichutz is a structure with countless rooms on each floor, each room
representing a unit of knowledge. False knowledge doesn’t exist in Pichutz, e.g. “the US
government is run by lizard men,” but can be brought in through factualization, e.g. “Barack
Obama said the US government is run by lizard men.” Lit rooms are acquired knowledge, and
dark rooms uncharted knowledge. Once a room is lit, it’s globally lit, but locally dark to those
who haven’t lit it. Since human knowledge is built mostly bottom up, the top of Pichutz is mostly
When you try to discover new knowledge, you launch yourself into the dark. Then you either
land into a dark room (come up with something right) or outside of Pichutz (come up with
something wrong). If the room you land into is globally dark (a discovery), you’re free to
command benefit from it, and choose to let it be globally or locally lit. But if that room is locally
dark (a secret), which means someone has been there before, you may be pushed out of Pichutz
to death. For insights into the dark top of the Pyramid, please refer to the Dark section after
going through this whole document.
Each room is associated with a probability distribution of probability values, which indicates the
percentage of users who have rated each probability value. I decided to use the probability
distribution of probabilities because it reflects the probabilistic nature of reality, tells you more
information than a probability
, and is the closest expression to a wave function, which has
later implications for Pichutz.
III. THE 1ST
It all started when I realized an apparent hierarchy between the conscious levels of a naïve
physicist and a wise politician (both are smart). What did the physicist know? Complex technical
concepts in physics and engineering. What did the politician know? Economic insights and
political secrets. What was perceived at the level of the latter’s understanding (e.g. the unified
axis of government and corporate power) might supersede the effects of both the former’s works
in physics and the very physical systems the former was studying or building (e.g. rocket
engines), e.g. by slashing state funding to his research. Conversely, a breakthrough in the
former’s field (e.g. E=mc2
, atomic bomb, hydraulic franking, any disruptive/transformational
technology such as mass production of robotic soldiers, directed-energy weapons, and portable
fusion reactors) might turn the whole political system that the later knew upside down. Thus they
seemed to be situated at two extremities of equal importance. If the physicist was “deep,” then
the politician should be “high.”
This hierarchy reemerged during my academic research when I realized that biological processes
emerge from chemical reactions, which can be derived from physical laws, which are described
and predicted by mathematical models, which are deduced from logics and founded on proofs,
which can be traced back to self-evident or assumed statements known as axioms. Economic
principles, political rules, and life hacks are all based on these scientific concepts to various
degrees, albeit unfathomably remotely. In this hierarchy of concepts, one is founded upon others.
This is basically the same hierarchy of the brain’s pattern recognition modules described in
[Kurzweil 2012] as functional units to recognize components of human languages—from
phonemes down to frequency bands of sound (perceived as pitches), up to words, phrases,
sentences, contexts, to abstractions such as attractiveness, irony, happiness, frustration, envy,
sarcasm, humor, nuance, metaphors, similes, implications, secret code, and so on.
Therefore, I have unified these compatible hierarchies into the 1st
(Conceptual) dimension. It’s
worth pointing out that we humans do not start at the top digging down or at the bottom climbing
up. We are born into the knowledge somewhere in the middle of Pichutz, where I call the Free-
, before embarking in two opposite directions simultaneously. The knowledge in
this Zone is the easiest for humans to acquire, and thus can be described as either Low or
Shallow. From the bottom to the top of Pichutz, the relative position of a room can be described
in order as Deep, Shallow, Low, and High.
Most of the time when someone asked me “Are you in deep thoughts?” I usually responded “No,
I’m in High thoughts.”
Skip to The 2nd
1. Higher rooms are founded upon Lower rooms. A Lower room would prop up (be part
of/produce a subset of solutions of/contribute to/prove/deduce) a Higher room, while the Higher
room would backprop (predict/contain/incorporate/dictate) a Lower room. The direct relation
between a pair of rooms is either a prop or backprop, which indicates their relative position along
dimension. For example, computing technologies for a speed of 1019
nondestructive brain scanning technologies are required to prop the mind uploading technology;
advanced technologies in propellant, engine, materials, design, pilot safety, tires, sensors, space,
etc. are required to prop the blueprint of a spaceship with recyclable rocket engines.
2. Between two rooms with no direct relation on different floors, the Higher room often
supersede/happens regardless of the Lower room, while the Lower room makes happen/possible
the concept of the Higher room
. In definition, a strategy is Higher than a tactic, and the latter
usually Deeper than the former.
3. Normally, a Higher room must be built upon Lower rooms. However, in fact, Nature
(evolution) has gotten ahead of humans since the Earth’s birth and lit High rooms long before we
could reach them methodologically bottom-up. These rooms called Hints provide suboptimal but
effective solutions for now, from which principles to develop optimal solutions can be derived
. Taking advantages of Hints thus accelerates humans’ exploration of knowledge.
Reverse-engineering the brain to model its interactions at the molecular level (or even quantum
level) and search for the “secrets of the human brain” is a prime example of acquiring Hints.
4. Rooms from High to Deep correspond to a decreasing level of generalization, abstraction,
, an increasing level of specialization, specificity, precision
. For example,
there are many theories for the brain with resolutions ranging from the whole cortex (High), to
modules consisting of neurons, to neurons, to very detailed level of molecules in a neuron, to
quantum phenomena in the brain (Deep).
5. Rooms from Low to High or Shallow to Deep correspond to increasing difficulty and energy
level. A higher “energy level” means either more energy/matter (higher number of molecules
studied, higher temperature, or higher speed), more energy density (e.g. from a white dwarf to a
neutron star to a black hole), or more energy/time/resource/brainpower consumption (e.g. a cool
quantum computer chip costs a far more amount of energy, resources, and brainpower to invent
than that of a spectacular electron-spewing Tesla coil). In fact, the upper limit energies of
particle colliders dictate whatever final resolution humans are able to deliver in terms of
scientific and technological development. Every time the Large Hadron Collider raises its energy
limit, a barrage of discoveries will follow, Theory of Everything as the likely final destination.
6. The Higher room A is to room B, the more probable it is that room A can supersede/obsolete
. There are two notable degrees of High-ness:
a) Paradigm-High: Room A is paradigm-Higher than room B if they have no direct
relationship but room A completely supersede/obsolete room B. For example, technological
paradigms were characterized by these technologies—electromechanical computers, relays,
vacuum tubes, transistors, integrated circuits—and next comes the self-configuring 3D molecular
circuits (and then my speculated nonlocal circuits). In another example, Israel Intel shifted the
computer chip paradigms twice when upgrading the 486 chips to Pentium M chips, and then to
Core 2 Duo chips. Previous technologies don’t direct prop later technologies which are smaller,
faster, cheaper, more efficient, and based on whole classes of Deeper concepts (e.g. quantum
entanglement), but they made POSSIBLE the conception of the next paradigms’ technologies.
For example, without transistors to run a transitional generation of computers and create a wide
range of tools to build tools, humans could never have invented integrated circuits.
(b) Meta-High: Normally Room A is Higher than room B when it supersedes room B, but not
ALL rooms on room B’s floor/energy level. It is only meta-High when it can supersede all rooms
on par with and below room B
. In physics, a High theory may apply to a highly energetic
system with a stable energy level, but a meta-High one can apply to a more realistic system with
varied energy levels. In life, a High dictator can dictate a Low farmer but there’s a small chance
the farmer can backstab the dictator, unless the dictator is High enough to predict and counter
ALL possibilities regarding the farmer’s actions (including his interactions with the
environment). This has profound implications in the theory of weaknesses
7. Deepest rooms are the foundations for Highest rooms. A commonly observed phenomenon—
also an optimal knowledge acquisition strategy—is the bouncing between the Deep and the
. For example, a sort of rapid ricochet between abstract theory and down-and-dirty
practice is commonplace in startups. This is because only discoveries of Deep knowledge can
enable discoveries of High knowledge, and vice versa.
8. A High theory is a Symmetry of one or many approximation theories (Approxes) if that theory
unifies all these Approxes
. For example, M-theory is a theory in physics that unifies all
consistent versions of superstring theory, which arise as special limiting cases of M-theory.
Please refer to an incomplete list of related topics on the 1st
IV. THE 2nd
“You are too SOFT” was what my best friend told me whenever I immersed myself in romantic
literature, popular fiction, songwriting, poem, fashion10
, dancing, basketball, pick-up lines, love
songs, Korean drama, football, etc. just to impress and conquer the girls who had those interests.
He didn’t say so to make me less human. He said so because he knew me so well, that such
knowledge would largely distract me, waste my time, and not serve my purposes in life.
But what if my actions had purposes? What if I knew that those girls were either very rich or
somehow a specific link in my grand plan? What if I lived my life only to please everyone and
do everything needed (even laugh and cry) for the sole purpose of long-to-infinite-term survival
and evolution? That reminds me of the notion of HARD scientists, who are the practical, goal-
oriented, and not concerned with whether the Singularity is near or not, nor whether an AI is
“strong” or “weak,” nor what the bleep “consciousness” is.
Related topics on the 1st
(Conceptual) dimension http://qr.ae/qWJCq
Almost everything to know about wearing suits http://qr.ae/q9UQn
So, can I be even Harder? What if I could invent all Deep mathematics, all High theories and
algorithms, build a machine to run them, dominate the world, and then merge with that machine,
survive cosmic catastrophes, exponentially evolve, and eventually dominate the universe
(absorbing all matter and energy and know everything that can be known in this universe and all
parallel universes if there be). If that happened, then at some point, there would be no distinction
between me and the machine—no “human touch.” That’d be beyond Hard—SUPER HARD,
which reverberates the concept of the Singularity Superintelligence.
Therefore, I conceived the 2nd
dimension consisting of three levels of purposefulness—a Soft
shell, a Hard shell, and a Super-Hard (S-Hard) Core. It’s not meaningful to divide any further.
Skip to The 3rd
Soft Hard S-Hard
Knowledge for the sake of
psychologically keeping humans alive
by enriching our mental life and
allowing “meaningful” human
Knowledge for the sake of
deriving S-Hard knowledge
Knowledge for the sake
of survival and
Limitless knowledge created by
humans but not contradicting Hard
and S-Hard knowledge
Minimum amount of
knowledge for humans to
Minimum amount of
knowledge for a
computer to achieve
A practically limitless realm of
knowledge inherent to humans;
basically a collective body of virtually
any possible random knowledge
which humans who can see beauty in
every trivial detail would create if
given unlimited lifespan, time, and
resources; every single slightly
different piece of creative work;
appeal to humans’ senses
All knowledge a machine
needs in the context of human
existence, to serve humans,
pass the Turing test, act like a
human, gain humans’ trust, so
that humans can help the
machine become even
smarter, until the machine can
take over the process of
All knowledge that a
machine needs to
survive, evolve in any
“maximize the diversity
of possible futures”12
simply “dominate the
Has human touch No human touch
For mental survival For physical survival
Why do people care about and remember all imaginary details http://qr.ae/qWJCN
Examples of Soft knowledge vs. Hard knowledge:
Photography vs. brain imaging, love vs. brain on love, winning
Olympic Gold medal vs. sports, fine dining vs. nutrition, fashion vs.
armor, sex vs. cloning, makeup vs. cognitive augmentation, music
composing vs. sound engineering, virtual sex vs. virtual doctor, slice
of life vs. science fiction, leisure chatbot vs. expert system, emotional
vs. security robots, contemporary art vs. classical art,
recreational drugs vs. medicine, recreational mathematics vs.
competitive programming, graphic-intensive video games vs.
classical strategy games (scrabble, chess, backgammon, Go, Bridge),
Victoria’s Secret diamond-laced bra vs. diamond drill.
knowledge that any
species would need to
robotics, AI, energy,
Soft climbing: acquiring Higher
and Higher Soft knowledge.
Pros: humans believe that they can
discover grains of wisdom in epic
fantasy novels, technological
inspirations in science fiction, the
universe’s mysteries while
traveling the world, or Nature’s
laws in this paper.
Cons: knowledge acquired is low-
prob, incomprehensible to AIs.
Hard climbing: acquiring Higher and Higher Hard/S-Hard
knowledge, including two processes:
(1) Fast top-down process of acquiring Hard knowledge
thanks to a large amount of Hints provided by Nature
S-Hard knowledge is then derived from Hard knowledge.
(2) Slow bottom-up process of acquiring S-Hard
knowledge (mostly in logic, mathematics, physics, and
engineering), analogous to a walk into the darkness
because of a low amount of Hints and insufficiently
resource allocation (i.e. time, capital, material, brainpower)
due to “Soft drifting.”
Soft drifting: the
tendency and preference
of humans to acquire
Soft knowledge rather
than Harder knowledge.
Reasons: peace and
allows Soft climbing;
Hard climbing is boring,
Hard rushing: is accelerated Hard climbing under extraordinary
circumstances. Normally people acquire knowledge in all directions.
Only in extreme/life-or-death situations such as war13
or a startup
under resource constraints14
would the human instinct remind us of
the ultimate Hint on top of Pichutz to aim for—“I must survive.”
During such times, humans did everything it took to survive, namely
taking the direct course up the Core (e.g. the Mahattan project), and
consistently achieved paradigm-High breakthroughs
, which are
later adapted for a wide range of civil (Soft/Hard) uses, e.g.
SpaceX on the verge of failure http://qr.ae/qW30W
difficult, and risky (i.e.
invest too much time and
money in the wrong
idea); Soft learners are
driven by joy, happiness,
chemotherapies are offshoots of chemical-warfare agents.
Driven by instinct, fear and hope. Consumed less physical resources
and time (more efficiency) but more human resources (casualties15
e.g. in war smart people on both sides died. If during peace time
academics may undertake researches in all kinds of directions, during
wartime they all had to focus mostly on physics and engineering.
Soft appeal: Anything Soft always appeals more to
humans because our biological systems are not
optimized for continuous Hard-working. Imaginative
technologies and science in sci-fi/superhero
are more enjoyable to kids and
humans in general17
. Soft knowledge activates the
brain’s reward system, triggers positive emotions,
signaling that there’s no resource constraint
endangering our existence, makes us forget all
purposes. Honestly, had I been born in a rich family
with enough money to do anything I ever wanted, I
doubt I would be writing this right now.
Hard-hitting: When a technology has
touched the boundary between Hard and
Soft then any further development would
only make the technology less efficient
due to the presence of paradigm-Higher
technologies. Whenever this happens,
humans have officially advanced into the
next technological paradigm. Any
further development of older
technologies would be for aesthetic (e.g.
steampunk, antiques, and pixel art) or
other irrational human purposes.
Rooms from S-Hard to Soft correspond to an increasing disorder/entropy (rooms with minor
differences become increasingly undistinguishable), an increasing redundancy (vast amount of
variations of the same knowledge), a decreasing “density” (more dark rooms), and a decreasing
The purposefulness of Soft and Hard knowledge can be subjective (varies from person to person)
and constantly changing (depending on the context including resource availability). A Soft
knowledge such as fashion knowledge may become Hard when you need it to approach a
powerful fashionista. However, if there’s not enough time for Soft knowledge, you might have to
utilize your Hard knowledge to kidnap that person and force him to do what you want, all while
taking more risks than learning fashion.
Lack of investment in acquiring S-Hard knowledge (e.g. SpaceX’s $1 billion investment
compared to Whatsapp’s $22 billion acquisition) is attributable to the fact that financial
resources were still concentrated in the hands of people who don’t possess the amount of
knowledge required to comprehend existential risks and large-scale vulnerabilities. It’s wise
people like Elon Musk and Stephen Hawking’s responsibility to force High insights down the
throat of people who own enough resources to ensure our survival (not necessarily the masses).
DARPA’s project to teach STEM https://www.topcoder.com/case-studies/darpa/
In times of resource crises or foreseeable disasters, humans would automatically focus on
Hard/S-Hard knowledge for survival’s sake. However, in most cases, the majority of the world
wouldn’t reach the Conceptual level High enough to realize the crises or disasters they are facing
and fail to act accordingly. Despite the seemingly more important nature of Hard/S-Hard
knowledge, given the current mental state of humans, it is necessary to pursue all three kinds of
. For more examples, please refer to the Example section
“Soft,” “Hard,” and “S-Hard” indicate not only the degree of Purposefulness of knowledge but
also a person’s propensity for the respective body of knowledge, that person’s activities, lifestyle,
and mentality that constitute such propensity. In my opinion, the current world’s Hardest
celebrity is Elon Musk.
V. THE 3rd
The Scopic dimension is the easiest to understand because it only tells you which groups a room
belongs to, for example arithmetic belongs to mathematics. For simplicity, it is like a supply
chain—raw numerical data supply facts; these data and facts supply disciplinary knowledge,
which in turn supplies all kinds of group objects such as disciplines, and users’ Pyrofiles.
We adopt both a standardized disciplinary taxonomy (each group called taxon, plural taxa, has a
well-defined taxonomic rank) represented by a Taxonomic Tree, and user-generated arbitrary
groupings (each group called scopon, plural scopa, has a floating Scopic rank) represented by a
Scopic Tree. A scopon may contain a taxon or taxa but not vice versa. The Scopic rank of a
scopon indicates its relative scope/scale/size (the number of rooms it contains) and the highest
rank of the scopa it contains. The hierarchy in this dimension is very simple—higher ranked
scopa contain lower ranked scopa. The first four ranks—(1) data
, (2) facts, and (3)
disciplinary knowledge—are fixed while higher ranks float.
Rank-1 rooms (data) are stored in form of a description text, along with numbers, arrays,
matrixes, or links to datasets. Rank-2 rooms (facts) are stored in form of free texts, photos,
videos, blueprints, or other forms of media. Rank-3 rooms (disciplinary knowledge) are stored in
form of free text and LaTeX. Future development of the 4th
(Language) dimension will host
other versions of these rooms in rawer, machine-friendly languages18
Each room is linked to and fro only one node on the Taxonomic Tree, but many nodes on the
Scopic Tree. Each node on either tree stores (1) a free text, (2) a pointer to a definition room (e.g.
“Physics” room which stores a broad definition of physics), and (3) pointers to all grouped rooms
(e.g. all physics-related knowledge). Each node on the Scopic Tree can be a technology, an
academic paper, a solution to a problem, a submission to a competition, etc. linked to all of its
related knowledge. For example, a “2018 Bomber” Scopic node links to and fro a description
text and all scientific knowledge employed in the Bomber such as metamaterials, explosives, and
computational algorithms. Philosophy and religion are two interesting group objects19
The purpose of this dimension is to encourage users to (1) group distant knowledge to create new
(and Higher) knowledge
or identify problems in one field that have solution in another field
(e.g. by running a search for rooms that share a majority of props from data and facts), and (2)
support existing knowledge and facts with numerical data in order to calibrate their probabilities.
In other words, by mapping all knowledge on Pichutz, we can (1) systematically explore all of
the overlapping boundaries between our exponentially expanding frontiers of knowledge, and (2)
hook up experts in different industries, enable them to look beyond just their narrow discipline,
juxtapose their knowledge next to each other, and identify the same concepts being described in
different terminology. This is important because one (sub)-field may have actually solved a
problem that another (sub)-field20
was attempting to solve but it is not realized because these
fields use different terminology. Similar processes are known in mathematics as “abstraction”
VI. SECONDARY DIMENSIONS
Knowledge bases employ only ontology and/or taxonomy; Kurzweil’s hierarchical model of
mind has only the Conceptual dimension; none gives you a sense of direction. Never before have
the vague concepts of knowledge dimensions been defined, refined and put together into a single
coherent picture like Pichutz. Yet, I still want to make this picture more complete. There are
some dimensions I want to add to or combine with the three primary ones—the 4th
dimension, the 5th
(Technical) dimension, the 6th
(Applied) dimension, which are all slightly
parallel to the 1st
1. The Language dimension:
With this dimension, I want to introduce a completely new notion of “language”: If human
language is a collective of languages humans use to communicate with each other, then a
molecular language should be a language molecules use to communicate with each other.
Examples of languages in order of decreasing human-friendliness are human language (between
humans), programming language (between humans and machines), machine language (the lowest
level programming language with op codes and operand numbers), electronic/digital/binary
language (between electrons and electronic components), neuronal/electrochemical language
(“language of though” between neurons or the hypothetical neocortical patterns
Problems in one field solved by solutions in another field http://qr.ae/q9BMK
DNA/molecular language (between molecules), and the futuristic computronium/quantum
language (between computronium)
Languages can exist in more abstract form depending on the structure and substrate of the
communicating entities. For example, whenever a genetic algorithm (GA) successfully generates
a set of parameters which constitute a solution that can magically solve all problems at hand
perfectly (e.g. speech recognition), these parameters embody High and Deep knowledge in an
entirely new language, which is incomprehensible to humans and has to be studied to derive
whatever secrets they are holding to achieve perfect speech recognition. The number and
physical relations between these parameters are dependent of the substrate of the hardware. As
exotic computing architecture with smaller substrates or substrates with more accessible states
(e.g. qubit) become available, down to a maximum at quantum level, rawer forms of language
will be invented by the GA, embodying Higher and Deeper knowledge acquired through the
interaction inside the machine or between it and its environment. Eventually, following this
pattern, the SS will invent the rawest possible form of language that only it can understand.
When this dimension is introduced, users will be compelled to translate knowledge into more
machine-friendly languages (programming code, pseudo-code, or mathematical symbols) and, in
that process, identify inconsistencies in their own thinking. With this dimension, I also want to
raise an issue that humans’ instinctual attachment to human language might turn out to be a
barrier to pure knowledge or “true understanding” of knowledge, which is largely formless and
not intuitive. In fact, my neurons have their own language that only they can understand.
Evidently, even though my ideas for Pichutz are so fluid in my mind, I’m having a hard time
putting them down in this document.
2. The Insightful-Technical dimension: is important for new learners who want to start learning
the most readable, insightful texts of a concept before diving into less comprehensible, more
technical details. When studying a theory, you will be able to choose between reading texts
explaining it in an insightful manner and those explaining it in technical details.
3. The Theoretical-Applied (or Global-Local) dimension: Knowledge should also be arranged
according to the degree of relevance to the environment of the users.
4. Correlation with the 1st
dimensions: The closer it is to the Core and the bottom, the
more disciplines become more computational (e.g. computational social sciences), more
quantitative, Deeper, more universal or “global,” harder to express in human language and easier
to express in rawer languages.
Skip to Building the Pyramid
VII. TERMINOLOGY AND ABBREVIATION
1. Dimension: can be understood as one of the three axes of the triaxial spectrum of knowledge
2. Room: a unit of knowledge, which can be anything that can be known—a qubit, a datum, a
fact, an opinion, a citation, an object, an idea, an action, a thought, an event, a phenomenon, a
theory, a plan, a conspiracy, or virtually any information to predict exactly anything that may
happen the next moment, etc.
3. Low-prob and high-prob: short for low probability (<70%) and high probability (>=70%).
4. Conceptual floor or floor: a level in the 1st
5. Hard shell and Soft shell: two outermost levels in the 2nd
Super-Hard Core, S-Hard Core or Core: the innermost level in the 2nd
dimension. The Core
is not called the “Core” for nothing. It’s filled up with the least possible amount of knowledge
required to reach the top. In other words, it’s the shortest path to the top.
7. Scope rank or rank: a level in the 3rd
8. SP: Singularity Pyramid.
9. SS: Singularity Superintelligence.
10. Hints: provide suboptimal yet effective solutions (e.g. brain-inspired computer architecture)
from which principles to develop optimal solutions can be derived. Fields that study or take
advantage of Hints include biomimetics, bionics, swarm technologies and communications, bio-
inspired computing, human-inspired Big Data Access Strategies, etc.
11. Insight: a High, usually low-prob, room lit based on a large amount of Low rooms. In other
words, an Insight is a loosely founded speculative knowledge which, if turning out to be correct,
can be refined and developed into a complete theory. Very smart people have a propensity for
very High low-prob fringe scientific theory21
. Insights and Hints together often led to very
popular but scientifically unfounded theories and concepts, e.g. Feng Shui, Chakra, Ki, Yin Yang,
Chinese/papal/shamanic exoteric knowledge, parapsychology, and alternative medicine.
Consciousness: is defined, in this paper, as the brain’s equivalent of a computer’s random
access memory. It doesn’t colloquially refer to self-consciousness, metacognition, the ability to
think about one’s own thinking or other misleading, unfounded, and pseudoscientific notions of
“consciousness.” Knowing how oft-abused the word “consciousness” is, please strip your mind
of all definitions you have ever come across until now.
13. A person’s consciousness: a subset of one’s own knowledge (as humans cannot recall
everything we know at once and that’s why we often make the same mistakes all over gain);
one’s perspective is a chosen subset of her consciousness (e.g. “Let’s take a perspective of a
mathematician and then that of a computer scientist to approach this problem.”).
14. Conscious level: is the level of the Highest room of a person’s consciousness. A High person
is someone with relatively High conscious level.
15. Symmetry: a High theory that unifies several approximation theories (Approxes).
16. Approx (short of approximation): is one of the approximation theories of a Symmetry, often
a theory that is only accurate and useful within certain boundaries. For example, the Albert-and-
Rex model of the brain22
is a fun inspirational read (the most upvoted Quora post) and can be
somehow applied in everyday activities until you can acquire a degree in neuroscience to
develop more accurate theories.
17. Cred (short for credentials): indicates the weight of a user’s rating for a taxon. A High cred in
one field doesn’t correspond to a High cred in another field. High-profile people may have
credentials in certain areas but not in other areas. For example, Deepak Chopra can have a great
voice in heart surgery matters, but be banned from any conference on philosophy or physics.
18. Logic-set: If you prospective employers/lovers check out your Pyrofile, they will be most
concerned with not only your Highest and Deepest rooms but how you connect rooms. Each
prop/backprop you draw between rooms reflects your unique logic. In the future, one’s logic-set
(in form of a webgraph) may matter much more than one’s “mindset” or skillset.
19. Mini-Pichutz with respect to room A: is a subset of Pichutz with room A as its local top. It
includes all rooms upon which the top is founded.
VIII. BUILDING THE PYRAMID
[Suh 2009] has observed that Wikipedia growth has essentially plateaued, hence unsolicited
contributions from human volunteers may yield a limited amount of knowledge going forward.
However, I believe that such a limit was largely attributable to the two words “unsolicited” and
“volunteer.” What if we give contributors a mission and benefits? It’d be very likely that we can
harness the full human potential to crowdsource knowledge.
The mission can be to push the frontiers of knowledge towards the Singularity in each dimension,
or simply to survive; benefits is access to Pichutz’s analytic/predictive tools described later in
this document. I wish to rely on machine learning to rush up the Core but have to admit that we
still don’t have what it takes go full S-Hard and thus still have to depend on human curation.
Pichutz’s initial objective is NOT to amass information (like most knowledge bases did) but to
encourage people to first post their Insights
(assumptions, “pearls of wisdom”) and then
exhaustively and logically validate them (the way most people in logical professions such
lawyers, scientists, and engineers do on a regular basis) by gathering all information and
knowledge to support them. That way, they will light all the rooms they have mentally skipped
on their way rushing up, in order to get rid of unfounded thoughts, make unexpected discoveries,
expand their Insights into full theories, and create a strong foundation to go even Higher than
where they are standing.
Most people don't have the resolution to verify their own assumptions, which are usually High
Insights derived from disparate memories throughout their life and suddenly put together
precariously through a mysterious process inside the brain often called the “intuition.” It's
troublesome to dig out the specific memories that led to these assumptions. Even when they have
successfully dug them out, these memories may turn out to be so remotely/vaguely connected to
the assumptions at hand that they would feel too embarrassed to admit. It’s important to
overcome this mental inertia in order to accelerate towards the Singularity.
Content-wise, since Pichutz’s focus is different from that of most knowledge bases (Wikipedia,
Freebase, Knowledge Grapth, DBpedia, Knowledge Vault, etc.), facts such as Obama’s
birthplace, music albums, or anything not in English are the LEAST of our concern. Rooms on
rank 1 and 2 (data and facts) should be filled up with mostly statistics, measurements, or free
texts describing events, e.g. “On December 21, 2012, Barack Obama conspired in a bar.”
Skip to Application
There are 3 main elements: the database, the visualization, and the AI.
We may use MySQL database with a layer of memcached servers in front of the databases. In the
future, we can "scale up" as we grow to a machine with lots of cores and ram, plus a replica. This
relational data model is the right way to structure most of the data for an application like Pichutz
and for most user-generated content sites. Schemas allow the data to persist in a typed manner
across new versions of Pichutz as it's developed, serving as documentation, and prevent bugs.
SQL lets us move the computation to the data as necessary rather than having to fetch tons of
data and post-process it in the application everywhere. We may switch a new database when
someone implements a distributed relational database with relaxed semantics. I will have to
consult with experts on this.
The main information in each room is stored in form of free text (for the most part of Pichutz),
mathematical symbols, programming code and, in the future, other data types to accommodate
even rawer languages which I have not imagined yet. In addition, each room is associated with a
probability distribution of integer probabilities and a representative probability in percentage.
The initial value of each room’s probability needs to be set by deep evidence-based probability
rating algorithms, with reference to Watson’s Deep Q&A. This initial rating is multiplied with a
cred of 100% in the equation of the room’s representative probability, which at first will
overshadow most users’ ratings.
A data visualization tool can be built with Three.js, a comprehensive WebGL library. Only three
primary dimensions are visualized in form of a 3D pyramid. The user interface (resembling
and Google Earth) includes basic operations like add, edit, remove, prop, backprop,
endorse, suggest edit, search, etc. on the top and advanced, licensed analytic/predictive tools at
the bottom. Rooms can be opened in multiple resizable, movable windows for reading or editing.
Figure 2: Each object stores pointers to every object an arrow originating from it points to.
A user can (1) create a room (including a text, initial probability of accuracy, and Hardness) and
tag it with standardized Taxa or create new Scopa to tag it, (2) endorse a room with a probability,
(3) create a prop or backprop between room A and room B (a directional relation) along with the
probability of that relation, (4) create a tag and add rooms, Taxa, or Scopa to it, and (5) create a
black thread that links 2 rooms the user deems similar (e.g. “Bin laden died” and “Bin laden was
killed”), or a red thread that links 2 rooms the user deems opposite (e.g. “Bin laden is dead” and
“Bin laden is alive”).
The lowest tag ranks (Scopic ranks) are rank 1 for “data” and rank 2 for “facts.” When a new
Scopon is created, its rank is equal to the highest rank of the Scopon it contains plus 1.
WikiGalaxy the visualization of Wikipedia wiki.polyfra.me
A set of rooms is designated to be on floor 0. The floor of a new room A is equal to the higher
value between the floor of a room B which props room A plus 1 and the floor of a room C that
backprops room A minus 1. When there is confusion (resulting from users’ unfamiliarity with
the concepts of prop and backprop), for example when room C is Higher than room B, the floor
of room A is set to 0 and flagged for reconsideration.
The AI will be developed in later phases to perform tasks discussed below.
Figure 3: Data associated with each room/user and how they correspond to the visualization
1. Question attachment: Users can attach questions or links to questions at Q&A sites, such as
Quora and Stack Overflow, to and fro a room in order to facilitate knowledge validation (down),
exploration of same-level knowledge (horizontal), to development of Higher theories (up)
example, lists of unsolved problems in mathematics24
and computer science25
doors to access the Deepest floors (and, next, the Highest floors). This activity encourages
massive collaboration in solving hard problems while leveraging other platforms.
2. Conflict flagging: For each theories/facts/events, there always exist several conflicting
counterparts. By flagging these rooms in conflict, we can draw other users’ attention to resolve
the conflict, by proving that most, if not all, of them are low-prob, or by coming up with a
unifying theory/Insight if all of them are high-prob. Dr. Kurzweil used to do this
3. Pichutz backup: Only one state of Pichutz needs to be backed up because users can add or
adjust content but cannot remove anything. If you think a room has something wrong, repost that
room with the wrong thing omitted and suggest users related to the original room to endorse the
new room. Rapid changes of rooms related to topics on the news can be recorded with
considerably more temporal resolution for research purposes.
4. Actions: The role of Pichutz is to help you or a machine know what actions to perform, not to
sit idly thinking forever. Therefore, the knowledge in each room is prioritized to present a part of
or this whole sequence: conditions concepts definitions actions results. At
programming language level, they should be in form of functions or procedures.
5. Users have to avoid and improve absolute, incomplete, or excessive statements. For example,
“The less government intervention the better” and “To succeed, you must be merciless” are
absolute and incomplete. You should improve and replace them with “In fast growing markets or
where leaders are incompetent, the less government intervention the better; otherwise,
government tend to play in important role” and “To succeed, when mercy is valued by people
who are valuable to you more than your gain from being merciless, be merciful; otherwise, be
merciless.” The more what-ifs and details, the better. You may add statistics, thresholds,
mathematical models, etc. but they have to relate to the core point of the room; otherwise it will
become excessive and needs to be divided.
1. Pyrofile and credentials (cred): “You’re rooms26
; everything is just rooms.”
In Pichutz, you are your knowledge, either self-sourced or externally-sourced. Cred is a measure
of your credentials for a taxon (not scopon) in percentage (0% to 100%). It is calculated as the
average of all of your learn-rates (also in percentage) for each accounted room in that taxon. A
user’s learn-rate for a room is accounted in the calculation of that user’s cred for all taxa that
contain that room. For example, someone who has learned 90% of everything there is in
theoretical physics can have a cred in physics as high as 75%.
There are three ways you can raise your cred for a taxon (tentatively equally distributed): (1)
earn learn-rate for any room in that taxon or the whole taxon by taking courses of Pichutz-
How many rooms are you? http://qr.ae/qSkVI
affiliate institutions or by submitting evidence of your knowledge, (2) be endorsed by high-cred
users, or (3) have your contributed content endorsed by high-cred users.
Cred(user X, taxon Y) =
Learnrate: user X’s learn-rate for each room in taxon Y, in %
Roomcount1: the number of accounted rooms in taxon Y
EndorserY: the endorser’s cred under taxon Y, in %
EndorsedY (input): the value of cred endorsed by that endorser for taxon Y, in %
RP: the representative probability of each room in taxon Y, in %
Roomcount2: the number of accounted rooms submitted by user X in taxon Y
When a user earns a cred value of x% for a taxon X from another user, the learn-rates of all
rooms under taxon X are set to the highest value between their current values and x%.
Accounted rooms of user X are rooms with a representative probability that exceeds a certain
threshold (tentatively 70%). Only the learn-rates of accounted rooms of user X will be accounted
in that user’s cred for all taxa containing those rooms.
2. Room and representative probability (RP):
Each room is represented by a probability distribution of integer probability values from 0 to 100
(telling you the percentage of users who rated each value) and an RP. The RP of a room is
increased in two ways (tentatively, 50-50)—direct probability voting (users input a percentage
for the room’s probability) and weighted probabilities transferred from (back)propping rooms
(users input a percentage for how much they think room A is important/relevant to room B).
RP(room A) =
Userprob (input): the probability rated by a user X for room A, in %
Usercred: the cred of a user X for the nearest taxon of room A, in %
RP(room B): the representative probability of a (back)propping room B, in %
Userweight (input): the user-rated importance of a (back)propping room B to room A, in %
Cred1, cred2: the cred(s) of a user X for the nearest taxon of room A and the nearest taxon of
room B, respectively, in %
3. Order of computation:
There are many objects (rooms, users, taxa, scopa) and the values (RP, learn-rate, cred)
associated with these objects are highly interconnected. They are computed using a simple
iterative algorithm. Each iteration calculates the next set of values based on the resulting set of
values of the last iteration. To avoid a feedback loop, this continues until the values converge.
If a RP exceeds a certain threshold, its room becomes an accounted room, and the RP will be
accounted in the computation of all related rooms’ RPs and all related users’ learn-rates.
Only learn-rates are stored in the database; creds are not. A cred of user X for taxon Y will be
computed when user X endorses a room, user, or prop/backprop a room in that taxon.
Therefore, the order is as follows:
Each room’s RP is computed from RPs of all accounted rooms related to it, and the creds of
the nearest taxa of all related raters.
A cred of user X for taxon Y, when called, will be computed from the learn-rates of all
accounted rooms user X has in in taxon Y.
4. Cost factor: The path to the Singularity is not smooth. It’s not to choose simply the most
obvious routes but routes we can afford. That’s why we have to keep a good check on alternate
routes’ resource consumption in order to allocate resources accordingly. Thus I’m looking into
how to associate a cost factor with each room.
And then come the AI
In the future, an AI must be heavily developed to learn from humans’ interaction with Pichutz to
perform a variety of complicated tasks.
1. Stitching: Most of the time, users would prefer reading knowledge in Wikipedia-article style
rather than browsing rooms endlessly in a 3D pyramid. They need a software to stitch knowledge
together in meaningful ways according to a wide range of criteria (e.g. probability >= 70%,
within 10 floors of room A, Hard & S-Hard, rank 3) into well-organized 2D articles.
2. Analytics/prediction: This will be the main source of revenue. Refer to Application section.
3. Validating: automatically prop or backprop knowledge to calibrate its probability.
4. Critical-thinking: performs a continual background scan of all of the existing rooms, flagging
those in conflict.
5. Leveling-up: Upon identifying contradicting rooms, this AI would chain them together, and
attempt to create a room at a Higher floor that resolves the apparent contradiction by providing a
perspective that explains both ideas.
6. Questioning: If the AI failed to create a unifying room, it would generate open questions over
the chained rooms, or search for current questions on Q&A websites such as Quora and Stack
Overflow. Humans would then attempt to answer these questions or extract knowledge from
current answers on Q&A websites.
7. Crossing: An AI will identify open questions in every discipline and search for solutions in
other disparate areas of knowledge, or provide an Insight from an apparently disconnected field.
8. Translating: translates knowledge along the Language dimension from human language to
more machine-friendly languages, at least programming language.
For more understanding about the role of the AI27
to Pichutz, and their implications for the SS28
please refer to the footnote links.
a. Phase 1: Prototype
(1) A comprehensive hierarchical knowledge map to support systematic academic research.
(2) A platform to validate problem statements, explore existing solutions to similar problems,
map all knowledge required to solve the problems, predict and map must-have features of the
solution, generate hypotheses, gather academic materials or experimental data to support each
hypothesis, etc. in preparing academic paper, project proposal, or solution to grand challenges.
(3) A platform to verify intellectual leaders’ insights, and for intellectual point-to-point critique
where each opponent can lay out assumptions, counter the other’s assumptions, back own
assumptions with further assumptions or facts or data, trace these assumptions to the deepest
level possible, and so on. Opponents may adjust their assumptions according to new information
provided be each other, split the difference, and resolve the argument; or arrive at assumptions
which either side cannot or refuse to reason beyond for subjective reasons. This platform can be
used to resolve any argument between you and your opponents simply by using a tool to overlap
your Pyrofile and theirs to see where they meet.
b. Phase 2: Full version
(4) An educational hub for users to map knowledge, keep track of their learning, follow Pichutz-
derived principles to produce original knowledge, acquire certified knowledge from Pichutz-
affiliate education partners and build their comprehensive online credentials (Pyrofile). A
selective Board of Experts will rate (assign a cred value of x% to each taxon covered by)
massive open online courses (provided by Coursera, Udacity, edX, etc.), programs of educational
institutions or organizations. When a user graduates, her learn-rate for all rooms in each covered
and rated taxon will be set to the higher value between their current values and x% times the
user’s grade in percentage, e.g 100% for A, 80% for B, 0% for F and so on. This will be a major
application because, in no time, all of education will be online
Roles of AIs to Pichutz http://qr.ae/qeQcw
From Pichutz to the SS http://qr.ae/qeQ6r
(5) A tool to verify and address long-term danger/existential risks to regulate knowledge
acquisition (e.g. encourage people to study engineering rather than literature).
(6) Provider of a distributed platform (“Home Pichutz”) to privately organize personal
knowledge; crowdsource/groupsource personal, military, or political planning assistance;
identify weaknesses and plan courses of actions to contain these weaknesses. For example, you
can post “I should work for Google” and “I should work for Facebook” and have your friends
prop these two options with their knowledge. In military, we can map weapons, countermeasures,
counter-countermeasures, human resources, psychological weakness, geographical advantages,
public ignorance, etc. (our own weaknesses and the enemies’).
(7) A platform to crowdsource systematic codification of hierarchical knowledge. Users can post
knowledge in rawer, more machine-friendly languages than human language, such as
mathematical language, pseudo-code, programming language, or even rawer forms. These
versions of knowledge are used to accelerate machine learning.
(8) In the face of the a paralyzing speed of academic publications worldwide (5 publications per
minute by 2013), Pichutz can serve as a template for academic publications so that we easily
identify the relationships between ideas, terminology, and data within a publication, between
different publications, and verify even the most technical details in these publications ourselves.
I also envision a similar template for solution submission to science competitions, so that voters
can visually investigate and compare technical advantages between solutions rather than watch
videos that tell nothing and depend heavily on production values29
(9) A platform for AI experiments
on hierarchical and machine-friendly knowledge, which
opens the door to profitable analytic/predictive tools; also a source of problems for competitive
programming sites such as TopCoder and Kaggle. For example, a tool can monitor the collective
knowledge of a market to identify market gaps; if there is a gap between the standard of
appreciation for food, hygiene, movie, etc. of Vietnam and that of the rest of the world, you can
provide products at just above their current standards and charge as much as you want; in my
country, people love paying for crap. That tool can be applied to the business of art30
. At some
point, Pichutz will spawn a new breed of ultra-high-tech analysts called “Paralysts.”
(10) A complete understanding of Pichutz opens your mind to a wide range of everyday mental
Nokia Sensing XChallenge Finalists https://www.nokiasensingxchallengevoting.org/
The business of art http://qr.ae/qefSR
A list of random everyday applications http://qr.ae/qZ7nW
X. THEORY OF WEAKNESSES
Pichutz is scary, because there are countless rooms to light. If someone lights a room you haven’t,
it becomes your relative weakness to that person. No matter how High you are, you can’t always
rely on your Highest knowledge which supersede most Lower knowledge to solve every Lower
problem. Your plan/tool based on High knowledge may help you succeed most of time but it’s
possible to fail when you suddenly need a Low knowledge that you lack. Even the most
advanced systems these days have to be regularly debugged lest being hacked
. You can be
considered invincible if you own the only gun in the town, but anyone can defeat you with as
easy as a kick in the crotch unless you have expected that and trained properly.
Theoretically, a weakness has the potential to affect ALL rooms above it. In other words, it can
cause damage to all people who have lit any room above it. The relatively Deeper a weakness is,
the more damage it may cause, e.g. the Heartbleed bug that unfolded early this year32
. The one
who lit the room has the option to share it for little recognition and reward, or exploit it for more.
It’s easiest to think of weaknesses as vulnerabilities in computer sciences, which are most
documented. In the future, weaknesses in life sciences33
will be much more debilitating.
Global weaknesses are objective, knowledge lacked from the whole collective of human
knowledge (e.g. deep space, aliens, unknown existential risks) while local weaknesses are
subjective, knowledge lacked by an incident person or, in order word, ignorance. A person’s lack
of a certain knowledge is to that person what that person is to the mankind—a local weakness.
Our world is plagued with local weaknesses rather than global weaknesses.
Dark or exclusively-lit relatively-Deep room, bug, hole, loophole, gap, vulnerability,
shortcoming, blind spot, opening (in martial arts), uncertainty, ignorance, weakness—the lingo is
endless, but the concept is universal. Let’s call it “weakness.” Whenever a weakness is
discovered, resources are concentrated to fix it. That is the whole law behind crime and war—
humans target and/or fix others’ weaknesses for a profit.
If Pichutz is a living organism and resources its blood, then weaknesses are where if targeted
Pichutz would cave and its blood would flow in to mend the wound and fill in the gap.
Weaknesses are mostly exposed for exploitation/fixing under extreme/life-or-death conditions
such as war
. Nevertheless, they can be every bit and piece of information you miss when you
“let your guard down.” 
If you overestimated your security and didn't buy a good door, a thief
can break in. If you miss a piece of news about the company whose stocks you own due to lack
of connections, you lose money. If you find a hole in the law system, exploit it and make sure not
to be discovered, you can beat the system and get rich. If your neighbor is good at electrical
Hacking DNA http://www.theatlantic.com/magazine/archive/2012/11/hacking-the-presidents-dna/309147/
engineering, he can steal your electricity. If you don't check your children's DNA, you may end
up raising someone else's kids. If your IT system is imperfect, a hacker will find and exploit it or
extort you. If you are not good-looking and careless about your appearance, you can lose a job or
be bullied. If a person hates you and you don’t know, when a situation arises, he will do
everything within his abilities to harm you. If you depend on someone, your dependence is a
weakness. When I was an environmentalist, whenever one of my fellows vandalized or swore
before the press camera like an illiterate dropout, I would look worse in the public eyes. Such
person was my weakness. Every time a Vietnamese steals from a local grocery in a foreign
country, I would look worse in that country so they count as my weakness. If one of your
teammates is not good enough, greedy, emotional, has a habit, someone to die for, or any
weakness, that person’s weaknesses become your weaknesses that other teams will exploit.
The State of Israel is a perfect example of a brilliant people having their weaknesses exposed
continuously and having what it takes to fix them all. [Senor 2011] has all the examples. The
Yom Kippur war that happened on Yom Kippur, the holiest day in Judaism, reminded Israelis of
what’s more important to survival than praying, and led to the institution of the Talpiot program
to recruit the best of the best humans in all faculties in Israel to defend the nation and more. They
fixed their weaknesses (the holy day); their scientific and technological breakthroughs achieved
while Hard-rushing (in war) fluidly found their way into the global markets since no country
could have all conditions met to get that High. With those breakthroughs, their alumni and those
of Unit 8200—another Israeli miracle—found leading companies in the world such as NICE and
Check Point. US Army recently bought the Iron Dome from Israel34
—a technology that only a
country showered with rockets on a daily basis could develop. Besides war, it used to be a matter
of life-or-death for this nation to grow food on dessert. Now one of their leading companies
Netafim is exporting their agricultural technology to the whole world, their wastewater recycling
rate of 72% being the highest in the world.
However, how about inherent weaknesses? What if I’m smart and rich, but short and weak? Just
knowing that doesn’t improve anything. Doesn’t that mean there’s a permanent, no matter how
small, chance that one day I can be toyed around and killed in humiliation by a brainless brawny
thug? In order to ensure my absolute survival, I must become meta-High—mastering all
knowledge on all floors below my conscious level—to be smart and rich enough to make myself
big and strong. However, the meta-High concept should always be regarded as a goal rather than
a realistic description of any entity that may exist at all. Turning into an elite fighter capable of
dodging bullets or warding off whoever attempts on my life is not inefficient, considering how
limited my resources are and how many floors I have to master to be truly meta-High.
US Army bought Iron Dome http://www.israelnationalnews.com/News/News.aspx/186951
However, there’s no need to be desperate. Most powerful elites in world raise enough layers of
High protection around themselves to seal off their weakness which is their fragile biology. To
defeat them and survive, attackers must also master High floors to break away from the grasp of
their High protection, which diminished the attackers’ incentives to perform the attack at all. In
case of network vulnerabilities, the one to exploit weaknesses of the systems must be High
enough to succeed while maintain his anonymity. However, the High people in the society have
contained the risks of these potential weaknesses so well with their High knowledge that no one
who had the relatively Deep skills to exploit the weakness also had the High skills to escape
justice. There were countless instances of hackers, including members of the Anonymous
, getting caught for their miscalculations.
To further understand inherent weaknesses, refer to the Robocop (2014) case study36
What we have discussed above is not to zero in on apparent solutions, but to encourage you to
acknowledge and accept natural laws of knowledge and the inevitable future resulting from them.
Any knowledge/technology WILL be weaponized and put in bad use37
by criminals, terrorists,
and hackers as long as the weaknesses inherent to us as a species (emerging from both our brains
and biological bodies) exist as targets for entities with fewer weaknesses to exploit and benefit
from. It is ineffective to encourage people to develop technologies “responsibly.”38
Exponential technologies will entail exponential weaknesses as they guarantee exponentially
widening gaps and holes in our knowledge for criminals to exploit, just like a fast growing IT
system harboring more and more vulnerabilities for hackers to exploit39
. Developing Higher
power (authority, money), technology (Internet of Things, synbio), and conscious level
(realization of profit from destruction) will only add more dark floors below—weaknesses.
Only with a powerful information tool such as Pichutz can we ascend High enough to come up
with not just strategies to address each threat separately, but even much Higher plan to seal off
our weaknesses completely.
Pichutz as a game
Why do I have to play Pichutz? Can’t I just learn instead of competing? It’s up to you. However,
over time, Pichutz will resemble a virtual reality rather than an “online knowledge mapping
system.” Like in any reality, competition is inevitable, as implied above.
Arrests of the Anonymous https://en.wikipedia.org/wiki/Anonymous_(group)#Arrests_and_trials
Robocop (2014) case study http://qr.ae/qZ8e7
Exponential crime http://www.ted.com/talks/marc_goodman_a_vision_of_crimes_in_the_future
Build the future “responsibly” http://singularityhub.com/2014/11/20/summit-europe-artificial-intelligence-
Massive bank attack http://money.cnn.com/2014/08/28/technology/security/bank-hack/index.html
The bottom line is: inside Pichutz, it contains an incomplete theory about itself.40
So, if you use
Pichutz but there're still things you do not know about it, they will become your weaknesses.
This will remind you forever that you reek of weaknesses and had better hurry up get High!
The Higher you get, the more you will feel like playing a game with your everything at stakes.
You interactions with knowledge at such levels will bring about enormous consequences. And
your Pyrofile will be one of your primary weapons when we introduce a feature that lets you set
the “conditional visibility” of separate parts of your Pyrofile. It means only users who meet
certain conditions can see the parts you want to hide.
Also, it’s worth noting that the moment High dark rooms (dark secret of the government,
military, and corporations) are open to you and the public, their first explorers—the Highest
masterminds will have moved up to the next floors. So it is an endless race.
But don’t worry! My goal in proposing Pichutz is to leverage the power of knowing and
empower you, giving you another tool to close the gap between yourself and those near the top.
So the bottom line is Pichutz will provide a paradigm-High knowledge service and learning
experience compared to Wikipedia. The chance is for everyone, but the reward is not.
XI. MY NEXT STEP
I believe Pichutz concept is powerful enough and its implications profound enough to attract
people with complementary skills to join me. My plan is to seek feedback to refine the concept,
gather like-minded partners, seek technical advices and endorsement, apply to incubators,
acquire must-have technical knowledge as fast as possible to communicate with prospective
technical partners and shape the vision of Pitchutz in a programming-intensive future, go to an
incubator this year, look for co-founders and work on this project full-time with the intention of a
I have envisaged a prototype for when I can go to an incubator, I will offer it to all participants as
a tool to map their highest-level insights in order to find like-minded partners and form teams;
resolve any conflict by exhaustively verifying their assumptions; list and verify problem
statements or existential risks they are tackling with facts, data, and disciplinary knowledge;
compare these problems to determine the most pressing; trace the root of current problems to
tackle it now, or prepare to tackle it later before solutions to current problems become
obsoleted; map knowledge from all fields required for their solutions (neuroscience, synbio, AI,
etc.) to identify which knowledge to acquire next and track their learning progress; devise
strategies to outsmart competitors, customers, and the public; identify where outputs of team
projects overlap to create synergy; etc. That would be a win-win for both me and those teams.
The consciousness of Pichutz http://qr.ae/qSmOI
As mentioned, I'm not just proposing a new structure for a knowledge system, but also a new
way of filling that system with knowledge. In order to establish the posting standards and input
the first substantial patch of knowledge into this platform as reference, I’m in need of a highly
selective audience, who is mature, wise, purpose-driven, multidisciplinary, and whose
knowledge covers a wide range from the deepest-level technical techniques to the highest-level
philosophical insights, and from the most pressing knowledge that contributes directly to the
survival of mankind such as deep learning or rocket science to the least likely knowledge to be
fed to a computer such as French fine-dining or Chinese poetry. Therefore, the incubator I’m
going to should attract an eclectic group for me to crowdsource my project’s critical database.
This project and another one are my life. Except for technical knowledge, I have a firm grasp of
the concept, its implications, and its “real” plan for the future—the future that I want to live in. If
not for that, I wouldn’t live another second. To borrow Eric Schmidt’s and Jonathan Rosenberg’s
words in “How Google Works,” I want to bet on an unimaginable future, because such big bets
can sometimes be easier to achieve than small ones, since they attract the best people.
XII. INTERPRETATIONS & NOTES
Limitations of Wikipedia: users’ ratings are equally weighted; information is linearly sorted
with no meaningful order; has semantic structures but presents no big picture. Solutions: take
user credentials into account, and add three dimensions.
Limitations of Google: ratings are based on the page’s quality rather than the knowledge’s
probability; information is linearly sorted with no meaningful order; has the Knowledge Graph
but no big picture. Solutions: add three dimensions.
The Singularity is expected to be near the Top, but not literally at the Top. The Theory of
Everything should be comparatively closer to the Top. I doubt humans will survive to witness
what lies at the Top.
Opportunities in education:
Over the next five years or so, all of education—K-12, college, and graduate school—will be
online, in every major language. Companies like Google and others are working to bring very
low cost, or no cost, Internet to everyone in the world. In the future, we will see very high quality
education, both in terms of online education with human thinkers, as well as computer-assisted
instruction by artificial intelligence, thus providing higher quality education to everyone. There
will need an educational hut to arrange that huge amount of high quality education in one place
so that users can record and share their life of learning in a single place41
An incomplete list of my questions in life in four main sets:
1. The structure of knowledge: How can I put all knowledge in the same picture? The whole
picture comprises of all fields. People are going up and down without anticipating where to
meet halfway in order to hold hands and walk up to the Top together. Why didn’t high school
teachers teach mathematics, physics, chemistry, and biology as a seamless body of science
where biology emerges from chemistry, which in turn emerges from physics, while
mathematics is the seed of science that describes everything? In theoretical physics, why
didn’t scientists from two camps (string theorists who go bottom up and model builders who
go top down) work hand in hand trying to meet somewhere in the middle? If scattered in the
body of religious knowledge to date are tiny pearls of truth/real knowledge then how should
they be positioned compared to scientific knowledge? Should scientists take leaps of faith for
unlikely breakthroughs? What does it take to draw a roadmap to the Singularity?
2. The map of a humans’ consciousness:
Why can’t big minds agree on everything with each
other? Why are some of the most intelligent people I know in their fields so naïve and
ignorant in worldly matters, and not conscious at all of the hidden danger threatening our
lives every second? How can I convince very intelligent people who have huge ego to sit
down and point out their elusive deficiencies in order to work out real solutions together?
How can I convince people who love or hate Hitler or other monstrous historical figures that
they were inevitable existence born out of the context of history rather than good or bad
people? How can I prove the relative and probabilistic nature of the reality (nothing is
absolutely true and everything is governed by a probabilistic distribution)? Do levels of
consciousness exist (People of higher level of consciousness are exposed to certain
knowledge that enables them to realize things closer to truth or on a larger physical scale.)
and how can I visualize someone’s consciousness? How can I really put myself in someone’s
shoes to understand why that person’s view is so glaringly contrasting to mine?
3. The reason for variety: Why are people still enjoying 2D games (Flappy Bird) while there are
so many awesome 3D games? Why not create the best thing but the best things? Why can’t I
have a one-size-fits-all all-time strategy? Why do people suddenly fall in love with retro-
styled music or fashion? Why do we still have martial arts while we have weapons? What
exactly do the arts of gardening, dancing, gourmet cooking, fine dining, fashion, photography,
poetry, etc. contribute to the humans’ progress towards the Singularity? Why are there
always people who over-appreciate particular inane forms of art (Mark Rothko’s black-on-
black—abstract expressionism—or Kasimir Malevich’s white-on-white—suprematism) and
A Q&A with Ray Kurzweil http://roi.ted.com/2014/10/15/singularity-university-the-future-of-education-and-the-
junk teen bestsellers (Beth Reekles’s The Kissing Booth and Maya Van Wagenen’s Vintage
Wisdom for a Modern Geek)? Why did junk teen novelists like Beth Reekles and Maya Van
Wagenen become two among The 16 Most Influential Teens of 2013? How could Nikita
Singh become such a best-selling Indian novelist if her best known works were named
“Love@Facebook” (2011), “Accidentally in Love…With Him? Again?” (2011), “If It's Not
Forever...It's Not Love” (2012), “The Promise” (2012), “Someone Like You” (2013), “The
Unreasonable Fellows” (2013), “Right Here Right Now” (2014)?
4. The nature of war and peace: Why does war exist? Does absolute peace exist? Why do
people fight? Are all kinds of violence and crime (robbery, murder, protest, IT hacking, etc.)
during peace time just another manifestation of war? Why do we need red team exercises in
computer science and life sciences? As long as there is profit in war, can there really be no
more war? Why were there many scientific breakthroughs in war that lead to the invention of
the computer and the Internet? What are the benefits of war? What exactly did wars
contribute to the collective sum of human knowledge? How did the process of developing
weapons and defense systems against certain weapons and uncertainty contribute to the
collective sum of human knowledge? Isn’t it true that every time a hacker found and
exploited a vulnerability for profit, the vulnerability will be fixed and the IT system will get
one step closer to perfection? Are all kinds of damage and losses the costs of collective
improvement and perfection?
Free services: research (able to keep track of frequent content updates), data mining, AI testing
(pay if discontinued due to low content endorsement rate), and Home Pichutz for
Paid services: advanced analytic/predictive tools (document parsing, market niche brute force
search, like-minded partner search), result sharing from AI developers, education (affiliate
institution rating), assess to hidden Pyrofiles (paying Pyrofiles’ owners who are willing to sell
their content), headhunting (candidate search, assessment), job analysis (based on current
employees), media (paid by journalists and news outlets who use Pichutz format), etc.
Wisdom and Intelligence:
One thing in RPG games that I think reflects real life is the separation of a character's two
attributes Intelligence and Wisdom. I have met very intelligent people who are experts in their
fields but naive and ignorant in most worldly matter. They could make instant calculation or
draw very complicated blueprints of buildings and machineries but couldn't understand why the
Titanic sank or why 11/9 happened or how disaster capitalism worked.
Many intellectual elders in my country think that Americans simply waged wars in our country
out of monetary greed. In most cases, high Wisdom requires high Intelligence, but not vice versa.
It’s easier to imagine an entity with extreme intelligent rather than on with extreme wisdom. For
example, the “ultra-intelligent” character Schön in Macroscope (1969) could play games and do
almost everything perfectly but never thought of living forever or dominating the universe. On
the other hand, the evil genius in Oliver Emberton’s to-be-published novel42
is comparably wiser
to advance his agenda and dominate the world, laying the foundation for his descendants to
achieve even more.
Deriving from important concepts in Pichutz, I have attempted to redefine Wisdom (which is
largely based on the Highest floor level one has reached) and Intelligence
Pichutz in comparison with Google Knowledge Vault: People as users are
evolving/interactive part of the picture rather than just static objects; presents a big picture; seek
High human subjective Insights rather than just objective knowledge; the purpose is not “ask me
anything”; knowledge creation is guided. If Google's mission is to organize the world's
information, Pichutz’s mission is to organize the collective body of knowledge.
The Deeper, the scarier:
American scientist Herbert A. Simon, in describing an AI program called EPAM, wrote in 1973
that to understand the mysterious EPAM program, he would provide us with three versions of the
explanation. One would be the form in which it was actually written (programming language)
with its whole structure of routines and subroutines. Alternatively, he could provide us with a
machine-language version of EPAM after the whole translation had been carried out. The third
would be the electromagnetic equations and boundary conditions that the computer, viewed as a
physical system, would have to obey while behaving as EPAM. That would be an acme of
reduction and incomprehensibility. [Kurzweil 2012]
The expert manager was itself a software program that was programmed to learn the strengths
and weaknesses of these different systems by examining their performance in real-world
situations. It was based on the notion that these strengths were orthogonal; that is, one system
would tend to be strong where another was weak. Indeed, the overall performance of the
combined systems with the trained expert manager in charge was far better than any of the
Bouncing between the Deep and the High:
High applications of mathematics in engineering and theoretical physics have driven much
development in Deep mathematics. Conversely, a Deep achievement43
in mathematics, logic, or
algorithm can usher a cascade of High technological wonders. In fact, the Highest theories
Evil genius takes over the world http://qr.ae/qSLHD
Take over the world after proving P=NP http://qr.ae/qSLiG
mostly depended on invention of more mathematics44
and more mathematics is always needed to
develop Higher theories45
. Computer scientists regularly bounce between developing Deep
algorithms (for the software) and building faster computers/networks based on High technologies
(for the hardware). In far future, the development of Deep computronium language will optimize
calculations of High ToE equations.
One of my favorite bounce happened when I learned about the probabilistic nature of reality
through the equations of wave functions and Schrödinger's equation (Deep). Thereafter, I only
describe things as “probable” rather than “accurate.” I also realized that if wave functions really
never collapse then the reality that we perceive is merely the interference between all the wave of
every quantum in our brain and body with that of every quantum in the universe. I found an
allegory in Young’s double-lid experiment with photons, where the visible light bands are the
reality while the waves of the photons are always present but invisible. That means everything is
predetermined, which resonates Einstein’s quote “God doesn’t play dice.” Yes, he doesn’t play
dice because he has known the dice’s wave function and all of its possibilities! Hence my
original expansion of Einstein’s quote “God doesn’t play dice, because He sees all sides at once”
and the huge dice showing three sides with a dark top on the 1st
page of this paper.
Abstraction in mathematics is the process of extracting the underlying essence of a
mathematical concept, removing any dependence on real world objects with which it might
originally have been connected, and generalizing it so that it has wider applications or matching
among other abstract descriptions of equivalent phenomena. Two of the most highly abstract
areas of modern mathematics are category theory and model theory. One of the advantages of
abstraction is: It reveals deep connections between different areas of mathematics. Known results
in one area can suggest conjectures in a related area. Techniques and methods from one area can
be applied to prove results in a related area.
Why we need Soft, Hard, and S-Hard while S-Hard is all we needs to reach the top?
If a perfect AI magically sprang into existence, in no presence of humans, it would gather only
knowledge at the Core, which is all it needs to reach the top of the pyramid. Unfortunately, it has
to be created by humans, so it needs Hard knowledge to co-exist with humans. On the other hand,
humans are living organisms so, in order to stay alive and productive, we need all sorts of Soft
knowledge—entertainment, tasty food, fun, joy, love, fashion, touch, art, photography, fiction,
sex, reward, music, flower, bar, club, firework, emotion, party, movies/books that elaborate
trivial life philosophies (Mr. Nobody46
, Cloud Atlas47
), etc. As a result, humans have to fill up
Soft, Hard, and S-Hard knowledge simultaneously until the Core has filled up enough for an AI
to use it to fill up the rest of the Core.
Inventing mathematics http://qr.ae/qSLJe
Need more mathematics http://qr.ae/qS0NW
Mr. Nobody http://www.imdb.com/title/tt0485947/
Cloud Atlas http://www.imdb.com/title/tt1371111/
“According to the law of accelerating returns, by the end of this century we will be able to create
computation at the limits of what is possible, based on the laws of physics as applied to
computation. We call matter and every organized in this way “computronium,” which is vastly
more powerful pound per pound than the human brain. It will not just be raw computation but
will be infused with intelligent algorithms…. The potential of matter to compute is also governed
by a very small number, Planck’s constant 6.6x10-34
joule-seconds. This is the smallest scale at
which we can apply energy for computation. We obtain the theoretical limit of an object to
perform computation by dividing the total energy by Planck’s constant. Lloyd shows how the
potential computing capacity of a kilogram of matter equals pi times energy divided by Planck’s
constant. Since the energy is such a large number and Planck’s constant is so small, this equation
generates an extremely large number: about 5x1050
operations per second.” [Kurzweil 2012]
Quotes on the language of thought:
“I very rarely think words at all. A thought comes, and I may try to express it in words
afterwards.” – Albert Einstein.
“As we experience our own thoughts and memories, we “know” what they mean, but they do not
exist as readily explainable thoughts and recollections. If we want to share them with others, we
need to translate them in to language.” [Kurzweil 2012]
Why a probability distribution of probabilities (PDP) but not simply a probability distribution
or probability or just a Boolean value of true and false?
It’s simple—to provide more information. We could have used a probability distribution of
probability distributions of probability distributions of… Unfortunately, I don’t think most
people would find such level of complexity meaningful and useful. If a room has a tentative
representative probability of 50%, a PDP would suffice to help you distinguish between these
different cases: (1) everyone has the some opinions, e.g. everyone rated 50%, (2) everyone has a
different opinion, e.g. ratings are evenly distributed from 0% to 100%, and (3) there are two
opposite camps, e.g. half of the ratings are 0% and half 100%.
Why are all Conceptual levels needed (not just the Highest)?
“It would be extremely difficult to interpret what these higher level patterns meant without
actually copying all of the information at every level into our own cortex. Thus each pattern in
our neocortex is meaningful only in light of all the information carried in the levels below it…. if
we were able to detect the pattern activations in someone’s neocortex, we would still have little
idea what those pattern activations meant without also having access to the entire hierarchy of
patterns… below each activated pattern. That would pretty much require access to that person’s
entire neocortex.” [Kurzweil 2012]