Make Internet great again, with graph-reasoning

Summary (since attention is limited): Graph-reasoning has existed as a concept since the 60s, and are likely soon developed as a software (thanks to AI, crypto and new social entrepreneurs). It is bound to have as much impact on society as Internet itself has had the last two decades. Like all technology, it’s a double-edged sword (both beneficial and harmful) which requires it to be developed with well-being of humans & other species in mind. Developed in the right way, it can bring global peace, increase well-being of all species and enable safe exploration of space, biology & consciousness.

These are the same goals as what people thought of the Internet in the early 90s, which is why it’s even more important to ensure this type of innovation benefit us all. You might still wonder what graph-reasoning is, so basically it is about connecting all information in the world such as collective human knowledge, human emotions, and each individual’s own experience in this world & universe. In the best case scenario, this technology will assist decision-making, determine truth/falseness in claims, fight fake news, connect people, help people build great habits and enable you to create your own utopia.

Now for the 15-minute read, which I wrote to be easy-to-follow.

An old vision, not yet implemented

In 1965, Ted Nelson published a whitepaper with the abstract:
“The kinds of file structures required if we are to use the computer for personal files [..] are wholly different in character from those customary (used today) in business and scientific data processing.”

Here I am, 54 years later, and feel like I could start my white-paper with the same abstract as Ted Nelson, who’s today 81 years old and not yet seen his vision be realized.

Today’s computers/smart phones/Internet websites miss architecture for thoughtful concept-understanding, debate, group-alignment and much else.

This is a large part of the reason why many see Internet in a negative light. Facebook & YouTube addictive/extremist recommendations. Reinforced beliefs from Google searches. Instagram depression. Extreme-censorship in China. Chinese propaganda on open Internet.

The visionaries of the Internet in 80s, 90s and 00s didn’t see these environments emerging. Rather, many believed that the “semantic web” would develop. That ideas would be shared freely and that as such peace & democracy would go global… and so it will… but it’s a matter of time and innovation.

To once again quote Ted Nelson, a bit more technical this time (hang in there). The information architecture “need to provide the capacity for intricate and idiosyncratic arrangements, total modifiability, undecided alternatives, and thorough internal documentation.”

intricate = very complicated or detailed
idiosyncratic = unique
modifiability = ability to modify (edit)
(since Medium miss a (+) button for Definition next to difficult words)

In this white paper I will (to end the abstract of Ted Nelson) “explain why the problem is not simple, and why the solution (a file structure) must yet be very simple [and] finally, I want to explain the philosophical implications of this approach for information retrieval and data structure in a changing world”.

The best time to plant a tree was 20 years ago, second best is today

A common Swedish expression is “A loved child has many names”. So is true for graph-reasoning. I don’t even know what to tell people I’m working on. Semantic web. Semantic reasoning. Giant Global Graph (GGG). Argument maps. Endless concept maps. Canonical Debate. Complexity management. Associative reasoning. Pros/cons lists. Mind maps.

By now you might get the point. It has something to do with how information relates to each other. How one concept is born out of another concept. How all knowledge is connected. That all information is connected.

As explained in the intro, this has been spoken about since the early days of the Internet. Ted Nelson is also known for coining the term hyperlink (a web link, just like that one). Although the discussion, this digital collection of human knowledge hasn’t yet been achieved. We’ve made progress, but many lost sight in the progress. Graph-databases has come into the world, and without that software, Panama Papers would have big problems to identify all the money launderers. Recent years, great initatives has also emerged such as Canonical Debate, Kumu.io and Kialo. But the fact still remains, most of conumer-Internet (Google, Wikipedia, Facebook, YouTube etc) offers an Internet that is wide, but is without depth.

Commercial software for depth has been around a while. You are likely familiar with mind maps software (hundreds of these). Then the next step: 3d-mind-maps such as TheBrain, DevonThink, Thortspace, TopicScape. Also argument maps like ReasoningLab, DebateGraph.org, Kialo. And corporate AI’s like CYC.com or SAS.com. However, they’re far from improving society in the mass-scale that this technology can.

An example of TheBrain — a non-linear file management being 20 years old yet still the best consumer-software for organizing networks of knowledge — but far from adapted to every-day usage for mass-market.

ince big efforts have already been put into this with limited result, to achieve this we really need to think new. And we need to realize we might need newer technologies to make it happen. 2D-UX is comfortable, and adaptable to mass-market. But to do efficient work in an information-rich environment, it is likely this software will have to support 3D-environments as well. At the same time, improvements in NLP (AI for understanding text (Natural Language Processing)) may prove to be useful as well. Not to mention how tokenomics (blockchain economical theory) can help scaling. This is a whole-of-technology approach, to build the next generation of information software.

But here I am. Not an expert in anything. Not a coder. Throwing out buzzwords. Without any developer-partner. With a great vision. And a product roadmap for how to build this.

Use cases

This article focus on how everyday people will be affected by this technology. First off, let’s talk about you can make use of digitized human knowledge in your life.

Purpose in the 21st century

The first and foremost important innovation that can be enabled by graph-reasoning is individual autonomy. Finding a purpose and path in this complex world. Making it easier to build habits in a high-stimuli environment.

As the contemporary philosopher Naval puts it

The modern struggle: Lone individuals summoning inhuman willpower, fasting, meditating, and exercising, up against armies of scientists & statisticians weaponizing abundant food, screens, & medicine into junk food, clickbait news, infinite porn, endless games & addictive drugs.

Use case example: habit-tracking

When interacting with the habit-tracking app you would tell it things like how you feel (described in words (I’m satisfied after the exercise) or with numbers (feeling 8/10 grateful after meeting friend). Over time, you can map things you like (positive emotions like gratefulness, satisfaction) and things you don’t want in your life (confusion, anger, loneliness). With enough data, the habit-app will start to recognize patterns of what activities make you feel well.

This could help you overcome unwanted habits (as mentioned in Naval’s quote above) whilst increase your mental well-being, impactful productivity, your health, your personal relationships and more.

The technology to build a good habit-tracking app is already out there. What’s necessary is just good design and a privacy-focused organization building it.

Use case: Brainstorm and combine ideas into a compelling narrative

Narratives (stories/beliefs) are used everywhere. It’s a large part of human sense-making of the world.

Finding a compelling narrative that 1) fit your intention and 2) is relatable from your audience 3) preferably has as few as possible unintended implications.

Narratives are great for things such as writing an article. Such as pushing through a decision to your friends or in a company. To motivate yourself. To set the message for a political campaign.

China is doing this very well. For example, the 30 year anniversary of Tiananmen Square (Pro-Democracy protest in 1989) was heavily censored in China. They post on Twitter (to influence non-Chinese audience) that the censorship should be celebrated, as it helps avoid political turmoil.

With China’s central-planning they are able to construct & maintain narratives that benefit The Party, as “Censorship is good”

Whilst China has big resources of brainwashers to determine best narratives, in West very few of us understand the process they use to set these narratives that are later used in psychological warfare (PsyOps). In this global environment, our own narratives get shaped by their propaganda, and we have no clear way of reviewing narratives spread throughout society.

Cluster of labels

Things become particularly useful, but also complex as we break down text from the original source its from in an article, to see it only as a standalone item, whether it’s a single word, a paragraph, a section or an article.

As Ted Nelson said:

The information architecture “need to provide the capacity for intricate and idiosyncratic arrangements, total modifiability, undecided alternatives”.

Use case: product management

Another complex use case of this is for product management. For example on the aspect of roadmap, a PM (product manager) can use previous meeting-notes and his/her own ideas and map up different features, all as individual concepts, label these with expected resources to develop, label conditions (any condition, ex. technical requirements, required partnerships, statistical data (with variations)), The product manager can then use the individual concepts and map these to a timeline. Of course this whole process is also collaborative, so all affected agents, whether developers, external reviewers, users, politicians, business partners and else can have their say.

Example use case: Debunk narratives

Just as graph-reasoning can be used to find compelling narratives, it can also expose false narratives. An argument-map that looks at the “raw arguments”, the pros and cons, how the narrative relates to concepts and what those concepts actually are — can help to understand which narrative is most true to reality, and which actors (nations, corporations, individuals etc. are Actors in game-theory) benefit most from different narratives.

This one is the very important innovation, as it can bring the Sokrates-type of discussion, whereas we don’t know much, but we can know with high certainty that many things are true (ex. nuclear warfare would be devastating, genders have differences, climate change is man-made).

ChangeAView.com was a Subreddit that got so popular (750k members because people desperately urge for a debate-platform) that made their own website, with a few cool features (best out there) but much more innovation is needed to determine real consensus on topics.

The big impact: understanding the world

Graph-reasoning can enable probabilities of understanding the world. It can bring science to the true forefront of our understanding. Not just science as in reinforcing beliefs. Science as in being critical of each-other’s claims. Being highly critical on provided data. Ensuring replication in scientific studies.

Finding consensus on theories based on statistical data. Applying various theories to concepts. Mapping out how the world of human knowledge relates to each other. Everything from physics, maths, biology, economy, sociology.

Utilize the understanding to optimize well-being

From what lens should decisions be made? It can serve individuals, corporations, nations a lot to have a thoughtful understanding — and this can be very dangerous for the world — as knowledge is power and if used wrongly it can make it easier to manipulate humans. As such, building and ensuring this software is only used for improving wellness of all conscious beings is a good way of using this software in a positive-impact way.

It can look at all issues and ask the question what is the best way forward to optimize wellness of all conscious beings?
This framework can be used to answer questions like: How much tolerance should there be for political freedom (like in liberal democracies, but not in dictatorships)? How high tolerance for religious freedom? How much free speech?

The biggest sector can be revolutionized: public policy

Governments provide great support for our modern world. But setting public policy is today done through representative democracy or dictatorships. Neither of these really offer a way for everyday citizens give feedback and have their voices heard. This is where graph-reasoning can have most impact.

Graph-reasoning will make public policy fit for the 21st century. By asking questions to citizens (or if finding a privacy-friendly solution, collecting opinions during the everyday life of citizens) this will enable policy-makers to take into account the true sentiment of the population when setting public policies.

Graph-reasoning assumes nothing is 100% true

On the topic of knowing things and articulating them. Nothing we claim is certain. Nothing. Not even “I exist”. ¹ Thereby, nothing anyone says, especially on more abstract levels are fully knowable.

This doesn’t imply we should throw someone under the bus for articulating things we don’t believe in. But neither that we‘re not to be held responsible for our words. What matters is society evolve from intense-judgement to a state of accepting we all make mistakes. We believe in things that we later change our mind on. That’s how learning fundamentally works.

And expressing oneself is a large part of accumulating knowledge. In today’s society we have very little safe-space for articulating ourselves. In liberal democracies, we quickly get labeled. In China and other autocracies, one can get jailed for saying the wrong thing.

Since we all believe in false things, we need safe space

We therefore need to have a place in our world, a place on our Internet — to articulate ideas where we can both be held responsible for them, but also be safe in that our ideas won’t be held against us. Or else future generations won’t learn to be thoughtful. And our generation will live with tensions in our minds for not being able to express ourselves. Therefore, judge me not for holding controversial positions. Attempt to understand where I’m coming from. And use this same framework for all other people.

But to make it really easy to be non-judgemental, a debate platform that gives a good overview of different positions on hot topics, displaying the main pros/cons, showing actors benefit from this position, what’s the typical background of people holding this position etc. This way, we will see where people form their opinions from — how it makes sense based on their life history. The ambitions and narratives of dictator of China, Xi Jinping, will show how he has reasons to articulate what he do. Same with Mr. Donald Trump. Or Jordan Peterson. Or Sam Harris (my favorite philosopher). Graph-reasoning can enable thoughtful understanding of people’s beliefs, and assist to create a non-judgmental mindset for observers, opening up for harmony and collaboration.

¹ “I exist” isn’t 100% true as our existence could be a simulation. One may say “I experience existing” but it ain’t either 100% true because as you say it, it’s a reference to previous few seconds, a memory, and thereby subject to fallacies. Thereby we can’t make any single claim with 100% certainty.

Production of information is critical to a healthy information diet

Providing people better rewards for producing information through 1) improved UX for understanding one’s own past writing (one’s own thinking) and 2) enable seamless feedback from friends, family & public (as set by user’s privacy settings) and 3) allow user to use their writings to form an argument-map to better understand themselves and 4) offer personalized feedback on produced information based on graph-reasoning and 5) facilitating real-life meetups based on produced information and 6) having a sense of purpose as produced information learns a distributed well-being-AI that assist the graph-reasoning. Note-taking + knowledge-graph + feedback = thoughtful reasoning

Personal-graph + social-graph = fulfilling social life

Graph-reasoning will be able to first and foremost 1) properly track our lives (habits, goals, beliefs (possibly even emotions)) as well as 2) give personalized feedback based on knowledge-graph of how to achieve the goals we set and 3) make it easier to achieve goals through the tracking mentioned above, or by 3a) gamify habits and thereby make it fun to do irl goal-hunting as well as 3b) assist beneficial and enjoyable real life experiences such as 3b1) recommenend irl-people, events & actions based on an extensive social-graph.

I hope I didn’t lose you in the last 2 sections. They were dense, highly abstract. To give further explanation of all these concepts would need another UX.
UX = User Experience (such as (+) expand-buttons of concept-explanations)

Graph-reasoning can determine net sum (net positive/negative)

A bit more technical from here on. But still easy-to-read.

Decision-making is about predicting outcome. In essence, we calculate the net sum. What can bring me most well-being? Often it’s not that abstract, but its about specific emotions, money, work-environment, friendships, family-matters. But in the end, all these decisions sum up to improving well-being of oneself and those in one’s surroundings.

Net-sum for all involved is immensely complex to determine, yet even a simplified calculation can be very useful. Some questions can be answered very easily. “Would a full-out nuclear war be desirable?”. Given we do some reductions ex. only looking at individuals only on planet earth, it becomes apparent that large future values of consciousness & well-being may be eradicated. This is obvious common sense, we don’t want nuclear-war. How about democracy/dictatorship? Not so certain. It depends on tons of factors. How effective are the democracies? How cruel are the dictatorship? To go ahead, the question need to be more specific, and calculate net-sum in an isolated area, such as “Does Venezuela provide better well-being to its citizens as a democracy/dictatorship?”.

What complicates determining best outcome is 1) time 2) what agents (individuals, companies, countries, etc.) and how important is each?

Reducing these two factors to a single variation (one sole individual for the biological lifetime of that individual (ex my own life until I die)), decision-making becomes slightly more logically tangible. One even more simplified framework one can use (not using net sum, but only the lens of regret) is the model of Amazon founder Jeff Bezos: “Will I regret it when I am 80?”.

Using graph-reasoning to answer this “simple” question can actually prove to be very useful. However, be aware that even though the graph is reduced to a single individual lifetime basis, it is still very complex to answer “Will I regret this decision when I am 80?”.

There are so many arguments to take into account. Alternative outcomes. Lets have a look at it.

First we need to define regret. What is regret? What type of things do I regret? What affects what I regret? How may these factors (societal norms & narratives, friends & family-opinions, personal values) change until I’m 80?

Writing the arguments for this — and mapping them on a timeline with various variations of how regret would look like is actually tangible.

Now the specific decision: Will it give me less regret when I’m 80 if I quit my Wall Street job & start Amazon?

Here comes into account family, friends, money, health, personal values. The affecting factors are vast and so the amount of variations, essentially unlimited. Thereby, reducing the factors & possible outcomes to only key signals, ex. 100 factors — would create maximum a few million outcomes. With each factor mapped to the definition of regret that previously was done, the graph-reasoning software could now run the calculation and produce an output: “With 97% certainty, you’ll experience 74% less regret quitting Wall Street and starting Amazon, see more variations”.

It won’t be perfect, but as Yuval Noah Harari says:

“The recommendations may never be perfect, but they don’t have to be. They just have to be better, on average, than human beings.”

More on Yuval Noah Harari (author of Sapiens) in risks with graph-reasoning.

Technical comments (for everyday people)

Abstractions make it lightweight

The well-being framework can in fact answer all and any questions. But all the answers, every single one, will be immensely complex. That complexity puts grand pressure on 1) data input 2) human cognition 3) computational capability. For these reasons, some abstractions (simplifications) need to be made, as well as limiting the argument-graph to certain conditions and contexts.

This design-simplification will enable graph-reasoning to be made with economical computational-power fit for todays IT environments. It will also enable smoother UX for the human-interface of the graph-reasoning — as well as reduce the data requirement to make sufficient understanding of concepts.

In essence, the graph-reasoning doesn’t need to understand the whole world, just the key signals/concepts. The magic happens in the relation between concepts, but to make the system lightweight so it can run in todays IT environments, these relations will be far from complete, only containing key signals.

Risks

Technology should like medicine, disclose possible side-effects to its users. Here I present side-effects how graph-reasoning can harm mankind.

Mainly, it’s about a concept called black ball invention, something that’s so harmful and shared with the public, it can’t be undone and the technology becomes humanity’s doomsday.

However, let’s also not forget people have been protesting against industrialization since 1800s. And would you agree we have higher well-being than the average farmer in the 1800s? New technologies bring risk, but also bring benefits.

If developed without verified identity

Botnets and propaganda armies. False narratives in argument-maps. Like Twitter, Reddit today, but the impact of the narratives much stronger.

Used by actors harmful to global well-being

Probably China already has a graph-reasoning. Either way, this software will enable their central-planning even more efficiency. So goes for other interest-groups that care more about their own power than the well-being of themselves and others.

Increased polarization

It is possible the combined knowledge-social-graph won’t be able to bridge the differences between people, but rather build tribe-mentalities and more clearly point out the differences the people have with no way of overcoming them. An example of this can be those who don’t trust the reasoning-process.

Fuel warfare against existing power-structures

We have nuclear weapons. It’s a huge accomplishment of the war leaders last 70 years to not have blown up the planet ❤ . We need to ensure we don’t go to full-out nuclear war in the coming centuries as well. Graph-reasoning will bring more transparency, which might be threatening to those in power today.

Over-digitalization losing humane values

This would be a digital tool, because that’s where the technology exists today. It will be artificial. Not natural to the human environment, or that of any other conscious animals on our planet. Like our social-media have had negative effects on humans’ real lifes, so this might bring intellectual discourse to the digital domain, making people speak less to each other in-person. Therefore it requires an organization which is not operating in the attention-economy (building addictive products) to build this.

If working effectively: Losing individual autonomy

Yuval Noah Harari put this very well that:

When you start giving algorithms the authority to decide what to study, where to live, who to marry, who to vote for [..] the algorithm makes recommendations and it’s up to you to decide whether to follow the recommendation or not. In many cases people will follow the recommendations because they realize from experience the algorithms make better choices. The recommendations may never be perfect, but they don’t have to be. They just have to be better, on average, than human beings. That’s not impossible because human beings very often make terrible mistakes, even in the most important decisions of their lives. This is not a future scenario. Already we give algorithms authority to decide which movies to see and which books to buy. But the more you trust the algorithm, the more you lose the ability to make decisions yourself. After a couple of years of following the recommendations of Google Maps, you no longer have a gut instinct of where to go. You no longer know your city. So even though theoretically you still have authority, in practice this authority has been shifting to the algorithm.

This isn’t necessarily bad, we lose individual autonomy. But so we already have. Do you know how to survive alone in the forest? No. Will humans a few generations from now know how to make decision on their own? Maybe not. Maybe that will bring them more internal peace, which they can use to explore the spectrum of consciousness, or do mighty materialistic things like space exploration or biological upgrading.

If work effectively: fuel artificial general intelligence

Altho the benefits from an AGI would be tremendous, it is also very scary as it can be a black ball technology that figures out how to destroy humanity.

Keeping the debate going

To fully explain all the use cases, how the software would function, what the risks are and much else require more than a 15-minute article. Preferably it could be presented & discussed in a graph-interface, but as you now know, that software isn’t out there yet. So, until then, I’ll continue my contribution of describing how I see this ecosystem evolving and where we need to be extra thoughtful.

Til next time,
Erik