Official Website: http://leverageresearch.org/
Leverage Research quarterly newsletter: https://www.leverageresearch.org/updates
Leverage Research Twitter: https://twitter.com/leverageres
Repository of links to Leverage Research’s published research, institute communications, and related materials.
About Leverage Research
Leverage Research is an independent non-profit research institute dedicated to aiding responsible scientific advances which benefit society.
They have three main programs:
History of Science program — studying how successful fields developed in the past, particularly in the early stages, through in-depth case studies of important scientific discoveries.
Exploratory Psychology program — disseminating introspection research tools and methods developed from previous psychology research (conducted during 2011–2019) to external researchers to conduct their own investigations and test our hypotheses.
- Bottlenecks Program — supporting external efforts to build fields and break bottlenecks in science and technology through events and collaborations.
What is “Leverage 1.0” and “Leverage 2.0″?
The terms “Leverage 1.0” (2011-2019) and “Leverage 2.0″ (2019 onwards) are typically used to differentiate between two very different phases of Leverage’s history.
The distinction is described as follows in this report by Leverage Research:
From 2011 to 2019, Leverage Research led a sprawling research collaboration investigating a wide array of areas in the social sciences. The environment was informal and unstructured, and researchers were afforded a great degree of freedom in the selection of their projects. Over time, the collaboration came to include many projects, several organizations, and at its peak, almost 45 people. This research collaboration was often informally referred to as “Leverage” or “the ecosystem,” and more recently—given the need to distinguish it from Leverage Research today—“Leverage 1.0.”
In June 2019, the Leverage 1.0 research collaboration was disbanded, with the various projects and organizations involved instead coming to operate independently and many of the staff who had been involved being let go. Leverage Research, then with only a small number of staff, completely restructured to focus on studying and supporting early stage science and applying the lessons of its research to aid scientific advance. Today, Leverage Research has a more standard work environment with three main programs—History of Science, Exploratory Psychology, and Bottlenecks in Science and Technology—and all of our staff have defined roles within the organization.
This restructure was designed to take into account the lessons we had learned from Leverage 1.0 and address some of the issues that led to dissolution. These included moving towards a more centralized and professionalized work culture with delineated roles and responsibilities, clear reporting lines, set work hours, and standard practices for setting expectations and managing performance. It also involved a narrower organizational focus, separation between work and personal lives, and much greater public communication to help others to understand and engage with our work.
The distinction is also discussed in this 2019 EA Forum post, this 2021 LessWrong post and a draft FAQ Leverage has shared.
Are Leverage 1.0 and Leverage 2.0 different?
Yes. Leverage 1.0 was a decentralized collaboration of researchers all exploring a wide range of topics which by 2019 involved around 40 people across different organizations.
Today Leverage Research is a 5 person organization focused on studying and supporting early stage science.
Did staff have bad experiences during Leverage 1.0?
So far three former employees who were involved in the Leverage project pre-2019 who have written about their experiences:
Since each account contains unique perspectives and a lot of nuance, they are likely better read than summarized here.
Some LW users also commented experiences prevlev-anon, polyphony , and OlliPayne.
Following Zoe’s account, Oliver Habryka, Rob Bensinger and others posted $100k bounty originally for negative information but subsequently updated to include positive or negative information. Cathleen describes her reaction to this bounty in the section of her account entitled “Conflict with some EAs and Rationalists and the role they’re playing.”
Leverage Research conducted an inquiry into staff experiences of Leverage 1.0 following Zoe’s post and wrote up the findings including mistakes and lessons learned. The summary of staff experiences in the report is as follows:
The inquiry, led by Paradigm board member Matthew Fallshaw, contacted 45 former and current team members and interviewed 15, having several extensive discussions with many of them. Of those interviewed, almost all reported positive or strongly positive overall experiences. Many described having learned and grown a lot in an intellectually diverse and stimulating environment, developing valuable relationships, and having future opportunities as a result of their participation.
However, most of those interviewed noted that while their overall experience was positive and they would still choose to do it again, their time with Leverage 1.0 nonetheless contained some very difficult aspects and bad periods. Importantly, a small minority reported being unsure whether they would describe their time at Leverage 1.0 as negative or positive overall, having experienced many strong positives but also important negatives and ongoing difficulties they were still processing. It was also particularly significant to us that they reported having worked with other individuals whom they believe may have had overall negative experiences. Further, although interviewees disagreed with the overall characterization of Leverage 1.0 presented in Zoe’s account, most believed she was trying to honestly convey her own experience and that others had had experiences that were in some ways akin to what Zoe described.
Investigating people’s experiences of Leverage 1.0 further, we came to believe that approximately three to five of the around 43 members of Leverage 1.0 in question may have had very negative experiences overall, particularly towards the end of the project (around 2017 to 2019).