History of Leverage 1.0

Initial note — I’m writing this essay in stages and plan to revise it in response to feedback. If you have feedback or questions or comments, at this point please send them to my personal or organizational email, or DM me on LessWrong (my username is Geoff_Anders).

After the essay is done, I’ll move it to Medium.

Leverage Research: History of Leverage 1.0

Table of Contents

  1. Introduction — 2019/11/22
  2. Initial design — 2019/11/22
  3. Getting started — 2020/1/8
  4. Research paths
    • Argument Mapping
    • Belief Mapping
    • Goal Mapping
    • Hard Intelligence Amplification
    • Soft Intelligence Amplification
    • Artificial Intelligence
    • Process Analysis
    • Intellectual Process Mapping
    • Polyphasic Sleep
    • [etc]
  5. Governance
  6. Investigating the Mind
  7. Cultural Evolution
  8. Dissolution
  9. Lessons and Mistakes
  10. Next Steps
  11. Appendix
    • Timeline
    • Supporters
    • Stories and Pictures

1. Introduction

Over the last eight years, I ran a research collaboration centered around my research institute Leverage Research. Having reached a natural breaking point in June of this year, I have been planning to write up a description of the project.

I expect the project to be of interest to four main audiences:

Research project designers. Leverage 1.0 had a very unusual and specific design, intended to maximize our chances of making progress on our primary research goals. This may interesting to anyone who is designing a research project and wants to have an example to think about.

Ambitious world-improvers. One of the primary goals of Leverage 1.0 was to investigate whether individuals working together could make the world much better. We looked at many possible avenues for this. It is possible that we learned something that could be helpful for others who are seeking to direct their efforts.

Introspectors and self-improvers. While we investigated many topics, the primary focus of Leverage 1.0 was understanding the individual human mind and understanding the degree to which it was possible to help people improve themselves. We developed research methods that we believe will be helpful to the academic field of psychology and which we are preparing to distribute there, but of interest to more people will be the general outlines of what we came to believe about individuals, groups, and self-improvement.

Friends, families, nearby communities. For a variety of reasons, Leverage 1.0 ended up being run in semi-stealth mode. That plus the complexity of the project made the project very confusing for our friends and families, as well as spectators in the nearby communities. By describing the project and how it went, I can hopefully demystify the project to an important degree.

Others may be interested as well. The project was quite complex and there are many, many strands to unpack. In writing about it I will try to capture both the ethos of the project as well as many of the concrete details.

2. Initial design

In 2011, I started a group with some of my friends. We all wanted to improve the world and we thought we would have a better chance if we worked together.

As a group, we were concerned about many problems in the world — global poverty, oppressive political systems, dangerous technologies, declining mental health, to list a few. We didn’t know how to help with these problems, or whether we could. But it seemed like the problems weren’t on track to being solved, and we didn’t know that we couldn’t help. So our plan was to try.

Already though, we were faced with big questions. Which problem or problems should we focus on? Should we aim small or aim big? For whatever problem we selected, how would we figure out what we should do?

There was a further problem. Some of us were convinced that the world was in danger. There were specific dangers — nuclear war, pandemics, disaligned artificial general intelligence — as well as the more general argument that the technological power of humanity was increasing and that the wisdom of humanity was not keeping pace. Should we work on smaller problems when the entire world was threatened?

Our group disagreed. Some thought we should aim higher, some lower. Some wanted to focus on alleviating poverty, some wanted to focus on global threats. From my experience arguing with people, I knew that even a group as small as ours would not be able to reach rational consensus on these questions in a short timeframe. At the same time, it seemed like we would need many more people to help us — and to be able to think independently — if we were going to succeed.

For that reason, I designed our initial team structure in a way specifically designed to handle disagreement. Rather than agreeing on a plan, my idea was that we would all run projects in parallel. People could work on each other’s projects or on their own project. We would all provide each other with feedback and ideas and other help. We would make some effort to reach consensus in the short run, but in general we would expect the group to converge only over time, as we all learned from our own and each other’s efforts.

At the end of 2010, a few days before we officially launched the group, we had a text chat on Skype where I outlined the team structure I wanted to use. A few days later, we launched the organization. We didn’t have a name for the group yet, so I registered www.provisionalwebsite.com and we started calling ourselves the Provisional Team. At this point it was just six of us, located in different cities, with each of us having pledged to work a minimum of 10 hours a week on projects we thought would help improve the world. This was our starting point.

3. Getting started

Our first task was the acquisition of basic resources. We didn’t have money, a location, or much of a team.

How could we do this? I had never created a large, lasting collaboration before, and I knew that to do this we would have to offer something of value to potential collaborators that they could recognize as valuable.

For myself, I had spent many years developing my philosophical views, thinking about scientific and theoretical methodology, and developing and testing models of the human mind. I thought these things were really valuable, but I didn’t think the value would be visible for a while. I would also soon have a PhD in philosophy, but unfortunately credentials weren’t going to be much help here.

I concluded that I had three primary things to offer. First, my most recent model of the human mind, which I called Connection Theory, or CT. From conversation I knew that some people found it sufficiently interesting that a few might want to investigate it. Second, a huge plan document I had made, where I had tried to lay out what seemed to me to be the right steps for improving the world if one could understand epistemology and the mind well enough. Third, the idea of the organization itself, a platform where people could help each other pursue ambitious plans for improving the world.

This was how I got the first volunteers. When I approached the first prospective volunteer, Mike, I explained the plan and CT. When I approached the second volunteer, I pointed out that I had the plan, CT, and Mike. Each new volunteer picked a project, and soon we had a fleet of small projects. We became more and more attractive to join, and the initial joining cost was low — just a pledge to work on improving the world for a minimum of an average of 10 hours a week.

By the time we had 20 volunteers, we had people who wanted to study happiness, society, and intelligence amplification, people who wanted to design consensus tools and planning software, people planning to start five different related organizations, one person planning to make a ton of money to fund the rest of us, one person recruiting, someone planning a campaign to help increase organ donation rates in the US, two people (not myself) building systematic philosophies, and one person (myself) coordinating the relatively small degree of utopian chaos our joint aspirations added up to.

In the summer of 2011, some of us got together to try to figure out how to raise money. The answer: network to funders who we thought would understand our vision, explain the project to them, and ask them for money. In October, several of us briefly moved into my apartment (in South Easton, south of Boston) to playtest working together for real. The result: we thought we could do it. In November, we did our first pitch and got our first funding, from Jaan Tallinn, technical co-founder of Skype. We also selected our name: Leverage Research.

The last remaining challenge was to get a location. I selected my top picks from the group of volunteers and proposed that we co-locate. People would only move to major cities, and our budget was limited. We chose Brooklyn and got an apartment together. In the beginning of 2012, we moved in — the founding full-time team for Leverage Research: me, Jasen Murray, Mark Lee, and Oliver Carefull. Now we had a chance to see if we could do any of the things we thought we could do.

4. Research paths