Corpus - Beta v3

Intro

This book is for those that are experiencing the following 3 feelings:

1. There are many huge challenges (global peace, overcoming the loneliness crisis, curing aging, solving AGI alignment) the world faces, but for some reason it seems we are stuck on these problems in particular, and it feels like we can do better. This feels confusing and disorienting.
2. There are many ideologies, political parties and movements that claim to tackle some of these problems, but they are very loud, very extremist and are not only missing but actively denying important parts of the problems. This feels repulsive and like there must be another way.
3. You feel powerless/helpless to do anything about these problems, and so does everyone else around you. Which seems paradoxical, because surely we can’t all be helpless? This feels alienating and lonely.

We have had these same feelings, and have met many, many people, from all walks of life, that have too. They form a core part of the current zeitgeist.

We think it is important to elucidate what these feelings might be pointing at, and that in doing so, it will become quite obvious what needs to be done to address the true underlying problems.

Part 1: What is behind the 3 feelings?

Feelings keep track of patterns in reality. The patterns don’t always point to something real, and feelings can be mistaken in their magnitude or direction. But if a certain feeling keeps popping up, it means there is something in reality that keeps generating that feeling.

We have some guesses for what the 3 Feelings point to:

1. Over the Horizon Problems and Constructivism

There are many challenges and hopes that we as humans care deeply about. Eliminating wars, scarcity, injustice, and forced suffering; bringing more leisure, knowledge, justice, freedom, security, and beauty into the world; solving aging, climate change and AGI alignment.

For some problems, we basically know how to go about solving them, we have a process.

For example: Building a skyscraper is very hard, but if we wanted to do it, we would have the usual methods of project management, engineering, architecture, material science etc to build a plan to solve the problem. It might still be very expensive and require expert input, but there is a repeatable process by which we can generate a plan that will solve our problem.

For these problems, we have what we call Constructivist methods. Methods that can make plans that reliably, predictably lead to a solution to the problem.

But for so much of what we care so deeply about, we seem to be stuck. Our current methods—academia, legislation, startups, genius intuition, or just plain concentrated power—simply aren't equipped to address them.

If instead of building a skyscraper, we want to find out which new law to pass to best increase the happiness of a country’s population, that is much harder. There is no reliable, repeatable process. There is no constructive method to solve the problem. There might be non-constructive ones, such as random guessing and hoping you get lucky.

Call these "Over the Horizon" (OTH) problems: not impossible to solve, but can’t be reliably solved with our current approaches.

When dealing with these kinds of problems, we fail (and flail) a lot. Success becomes more due to sheer luck than good execution.

Our claim: OTH Problems are exactly the problems we feel like we are stuck on.

Most of the problems listed in the intro, which we care about, but there is no project to address, are exactly the problems for which we have no constructive method to address them. They are beyond the horizon of what our current constructive methods can handle.

But historically, we have often gone beyond our old methods and extended them into new domains beyond the horizon. We think there is a common core to this process of moving beyond the horizon, and that is aligned with truth finding and the scientific method.

We call this core Constructivism, and it is why we have the feeling that we can do better.

Despite what one might think, given that OTH problems are by definition “beyond our current means”, there are patterns and regularities to tackling OTH Problems.There are convergent bottlenecks and tools that can be used to address OTH Problems in general. We can build constructivist approaches to OTH problems.

Later in this book, we will talk about our best guess for how to build projects to solve OTH Problems.

2. Humanism

Looking across modern political, scientific and personal life, we don’t see any project seriously tackling these problems. Worse, almost all current ideologies and movements that claim to tackle such problems deny and reject major aspects of human values while doing so.

Communists reject individualism for economic equality. Libertarians reject moral duties for freedom. Degrowth rejects humans for nature. Techno-accelerationists reject the ends in favor of the means. And across the political spectrum, people defend morally indefensible actions of their leaders and allies.

Human values are inherently complex, multifaceted, and sometimes contradictory—and this is at the very core of what makes us human.

Humanism is the principle of grappling with all values of humanity, in all their complexity. And its absence in current ideologies is what we feel as repulsive.

We are currently lacking a Humanism for the 21st century, one that embraces the complexity of both our values and the problems we are facing, instead of grinding down the human soul to dust.

Later in this book, we will talk about how to build humanist methods of solving OTH Problems.

3. Coordination

When one wants to solve a problem in the world, often the most immediate plan is of the form “I want X. To get X, I need Power. So I will go get that Power, and fight for it if necessary.” This method is so ubiquitous, it is often seen as just the “logical default” and even “only option”. Call this the “Power Strategy.”

The Power Strategy is extremely negative sum, as people pursuing the plan, even in pursuit of the same or very similar X will compete with one another for scarce Power in destructive and wasteful competition.

But there is another option, the “Coordination Strategy.”

The Coordination Strategy is something like “I want X. I’m a human, so probably there are other people out there that also want X. How can I empower everyone in their pursuit of X, rather than just myself?”

This turns striving for X into a positive sum game.

Solving OTH Problems is often strongly bottlenecked by people choosing Power over Coordination.

There are many reasons for why people choose Power, sometimes out of malice, often not. Quite often, it is simply because there is no process to enact the Coordination Strategy together! It’s not that there is such a process and it is being used for evil, this process just does not exist!

This leads to the feeling that everyone would be better off working together, but for some reason cannot, and nothing happens and everyone feels powerless.

Later in this book, we will talk about how to build the methods to enact the Coordination Strategy together.

Part 2

We want a method that can solve OTH Problems in general. What criteria would such a method have to address?

Constructivism 101: Reliability

We have already mentioned the concept of Constructivism, which we think is at the core of not just science, but of how to extend science into new fields.

The unifying principle of Constructivism is: When we build something, it should work for the reasons we predicted it would work. There is an unbroken causal chain from our current state to the final output.

Our plan doesn’t have to be formal mathematics, but there should be a plan where every step is reliable.

Many plans do not have this shape. For example, many people trying to improve their countries just aim for “more growth/GDP”, thinking this will lead to more human flourishing, without having any reliable model for what connects those two steps, if anything. And indeed, despite the highest GDP ever, fertility has crashed, and they have no answer for why this is or what to do about it other than “get even more GDP.”

A plan is constructive precisely in so far as when it breaks, one can find which step it broke at, and debug from there.

To solve OTH problems, we need to make plans that actually reliably solve them, hence they must be constructivist.

Bootstrapping

Solving OTH problems, basically by definition, involves doing things we cannot currently do, or knowing things we don’t currently know. So we will have to acquire these capabilities.

There is a “core” to the process of acquiring new capabilities. We can call this core “general intelligence”, or “general problem solving ability” or whatever we want. Whatever we call it, it is what we need in order to acquire the capabilities to solve problems we can’t currently solve.

Bootstrapping is what we call the process of building this core/minimal set of tools and methods that lets us reliably acquire any missing necessary skills, habits or knowledge to address a given OTH problem.

Bootstrapping will to some degree necessarily be a clunky and difficult process, because it is itself an OTH problem. We do not have good, reliable methods for bootstrapping (yet).

Some of the processes of bootstrapping are simple. For example, there will be many facts you will need to know. But learning facts is only useful if you have sufficient long term memory to make use of them, so you must have a habit of taking notes and reviewing them regularly. Anything not written down will be forgotten. You don’t have to transcribe everything perfectly, just enough so you can reconstruct what it was you needed to remember. If you’re unsure what is important, write down more.

There are many habits you will need to build to tackle OTH problems. Most of the bottlenecks and processes to build such habits are far more about discipline than about intellect. As such, building the necessary minimum dedication and discipline is a must.

For this, you must carve out a pocket of time, establish a practice. This can just be 30 minutes a day where you take the time to reflect on your notes and concepts you want to internalize. You should spend this time in quiet, alone and taking notes as you think.

After you have the “bootstrapping core”, with which you can learn new habits and other things, our current best guess is there are about 10ish main similar practical bottlenecks you will need to develop reliable processes for from there. We will likely encounter more as we work on more OTH problems, and we therefore also need to be able to recognize when we have encountered a new bottleneck that is not on the list, and to bootstrap against it.

There is a lot more to be said about learning and bootstrapping. You can read some of our tips in the appendix (TODO: doesn't currently exist).

Focusing

Once we know what we need (a constructivist plan), and a process for gaining the skills we need to build and execute the plan (bootstrapping), we need to not lose focus.

Anything not written down will be forgotten, including your goals. Write down your goals, your plans, your thoughts, everything. Put all your reminders, meetings and plans into your calendar and follow it religiously. If you need to think of something at a later date, don’t leave it up to chance, set a reminder.

At all times, you should have a north star (written down, of course), of where you want to go, what you want to achieve, from which you backprop everything else. Of course, you can change your north star later, but never forget you have one, and schedule yourself to regularly review and remind yourself of it.

There are two kinds of mental phenomena that can be particularly pernicious in breaking your focus: Traumas and Compulsions. Traumas are things your thoughts are repelled by, Compulsions are things your thoughts are attracted towards.

Traumas prevent you from considering options and paths that might be important by preventing you from thinking through certain “forbidden” spaces. Compulsions also prevent you from seeing the whole space by drawing you towards a specific subset.

Noticing your traumas and compulsions, what things you like and dislike for un-endorsed reasons, is very important, not just so you can work on overcoming your blindspots, but also so you can inform others where you might have unusually many blindspots, so they can cover that space for you.

Another very common source of lost focus is Trolls. Trolls are people, groups, events and processes (such as “the news cycle”) that, for various reasons, optimize to draw your attention to somewhere useless. They might be particularly annoying, optimized to get under your skin and infuriate you. Or they might be particularly nerd-snipe-y, offering particularly fun rabbit holes that draw you in to waste your time. Both are dangerous.

Noticing trolls is an important skill to master. Ask yourself when you find yourself drawn to reacting to someone or something whether this is actually the most important thing you could be doing, and resist the temptation. With trolls, the only winning move is not to play.

Finally, the ur-source of lost focus is simply Entropy.

All things equal, order decays into chaos, focus decays into distraction. There are always other priorities in your life that will vie for your attention, projects that start out feeling exciting and novel will start feeling tedious and frustrating, and you will simply forget things. Even strong forms of memories decay. Notes become too complex and garbled, post-its disappear and get thrown into the trash can. Bits rot, books unravel.

Entropy is decay, and everything complex decays unless maintained. Entropy is not a failure on your part, it is not because of mistakes you are making, it’s all the maintenance costs that accumulate despite you doing everything right.

Constructivism 201: Efficiency

Most plans are non-constructivist. Of those that are constructivist, most are not efficient. By “efficient” we mean “could be feasible given resources.”

There are a number of general properties you will want to observe when making your plans. Some of the most important ones are:

• Always have a baseline plan.
• Once you have a baseline, iterate on it.
• Write down your plan.
• Ground every step of the process wherever and as much as feasibly possible
• Only consider the immediate next step. If you don’t have an immediate, concrete, grounded next step you are executing on, do not waste too much time thinking about later steps.
• Move your key bottlenecks close to the next step rather than late. Test your core assumptions early.
• Discard plans that have more than one sequential bottlenecks (i.e. more than one extraordinary thing has to happen in series for the plan to work)
• Favor parallelism wherever possible
• Apply the precautionary principle for wipeout risk. The world is non-ergodic, sometimes you die

Coordination

One of the most important sources of intractability is how everyone is fighting and hurting each other, even when they want the same things.

Failing to coordinate is as if your right hand and your left hand didn’t know what the other was doing, and fighting each other to pick up the object you want to lift, each knocking it out of the other's grip so they can pick it up.

Conversely, if you and other people are well coordinated, it should be like a coherent group or body that is winning the game.

Failure to coordinate is what happens by default. Coordination is a complex, delicate state that must be engineered deliberately. If you don’t create coordination, there will be no coordination.

So whatever the exact method is, one must address this failure mode, how to reliably prevent multiple actors (and even one 's own contradicting desires and emotions at times!) pursuing the same goals from shooting each other in the foot.

Part 3

Any general purpose methodology for tackling OTH Problems must address all of the above, at the very least.

Our best guess at how to build a humanist method for addressing OTH problems is by developing a reliable method to repeatably build groups of 5-20 people to work on OTH problems.

Our inability to tackle OTH Problems at the level of humanity is downstream of our inability to work systematically and collectively on the most important problems together and stay focused. Humanity has just simply not discovered a reliable method to do so consistently.

We think the focus should be on the group, rather than on individual rationality, because, in addition to avoiding dangerous psychological dynamics on the individual level, the group level is in fact generally the more agentic layer of human action at scale, and therefore more worthwhile to target. It is also generally cheaper, easier and safer to change groups than it is to change individuals in many ways, such as changing their vibe/personality, and for avoiding decaying into cult dynamics.

Ultimately, we want to build a method that lets us develop the memeplex or collection of norms/institutional processes that is coherent enough over time and robust enough to carry humanity into a good future. And we want to continually try our methods against that goal as we develop them.

Our current workstreams to develop such a method has three parts:

A: Theory

There is a certain amount of basic theory and vocabulary that needs to be established to be able to talk about and work on OTH problems. These basic concepts and descriptions of the problem need to be written down cleanly, so that people that want to come together have a base to build off of.

This is an intrinsic part of bootstrapping. Our first attempt is definitely not the Final Theory, our concepts and descriptions won’t be the perfect or final ones. But trying to address the problems we care about with the current common vocabulary/methods/theory alone would just be way too hard.

We need to build up at least a basic set of better concepts to work off of. Concepts like “Coordination Problem”, “Humanism”, “Karmic Thinking”, “Entropy” appear many, many times when addressing OTH problems, and we don’t want to have to constantly redefine them.

We’re not trying to build an entire new theoretic framework from scratch, just take a solid pareto improvement step towards better basic tools for us to work off of.

You are reading the output of this right now, The Corpus!

B: Community

There is a common pattern for ambitious projects and groups, where there is one or a few “founders” that fundamentally drive and provide the lifeblood of the project.

The humanist future that Torchbearer stands for should (indeed, has to) work without its founders. The bus number (the smallest number of people that would need to get hit by a bus simultaneously to destroy the project) is just way too low otherwise, especially for a project meant to scale across thousands or even millions of people and last for significant time. So the project must be distributed across people.

We also want the movement to have agency, to be able to accomplish things in the world collectively. This requires building an actual community, i.e. a group of connected people that can take actions together under a common group identity.

If we just have many individuals acting independently on these principles, we are back to individual rationality, rather than group coordination, necessitating each individual to improve themselves to the level of becoming a Founder as the only way to increase the bus number. For a community, it is much, much cheaper to increase the bus number, by introducing redundancy.

While improving individuals to Founder level all by themselves would of course be great, it is not something that has been bootstrapped yet. So we need a place to welcome and empower people to do things as part of a larger community.

A community also provides momentum. If we have an active community with a solid track record that one can trust, we won’t need to convince each individual that the content is good and the project worth contributing to. Without a clear brand and community, any good actions and outcomes will not accrue to the project.

A strong and productive community is central to the success of the project, but it’s important to keep in mind that most communities default to the lowest common denominator. We must invest the (significant) energy needed to have an actual humanist community.

C: Practice

Ultimately, we must actually figure out how to reliably, effectively and scalably create functional and morally aligned groups that tackle OTH problems.

This is not currently a scientific question. When something has become “science”, we can reason about it from our armchairs (or at least computers) and draw up plans that will reliably extrapolate to the real world. Here, that is not the case, we must constantly check all of our ideas and theories against reality.

So we need to do actual experimental science, experimenting with many many small groups in parallel with different norms, policies, protocols etc, to find processes that actually work and are replicable.

We need to form many small groups, communities, companies, etc, and iterate on various norms, protocols and systems until we find repeatable recipes of social organization. It should become possible for us to repeatedly and reliably take a project idea and spin up a stable and productive group of 2-20 people to tackle that problem without further supervision.

A very common failure mode for such groups is that they end up just talking. Many people like talking, few like acting. So we must filter strongly for acting, while still maintaining the high level of alignment needed such that the actions actually benefit the cause.

This is one of, if not the, core obstacles we face: Making acting tractable. It’s important that we can identify and directly experiment on such core bottlenecks whenever possible.

Even if groups die or have short lifespans, that can be ok. If the cost of creating them is low enough, we can just continually recreate them. But if they don’t act, then why are we even bothering? It’s nice and karmic to have people come together to talk about important things, but this is not a constructivist plan to solve OTH problems.

Once we can reliably make small groups that are well coordinated, this also opens the possibility of using similar methods to coordinate between these groups to produce even larger and efficient communities and movements. We expect building methods that can reliably scale to individual groups of say 1000 or more people to be too expensive. We think that instead, having many independent groups, or a kind of fractal hierarchy of “cells”, to be more tractable than trying to create a single massive group.