Note from Matt Fallshaw on Mon, Oct 18, 2021 on behalf of Leverage Research and Paradigm:

The below document is the original  Ecosystem Dissolution Information Arrangement made in 2019 and signed by 25 individuals (some with modifications). I have lightly edited it only to remove individual names of people and some pieces of intellectual property. It is otherwise exactly as per the original.  

I also asked Larissa to add comments using Google Docs’ comment feature to clarify the organisations' current perspective.  

Information arrangement — post-ecosystem dissolution

Purposes

• To help us manage our PR going forward.[a]

• To diminish risks from problematic [b]spread of information.

• To give individuals and organizations control over their own information[c].

• To minimally restrict information spread otherwise.[d]

Specifics

(1) General truce

• Be generally positive[e] about each other

• Don’t talk smack, etc.

• Support each other’s preferred (~true) narratives

(2) Expect a positive future story

• At some point,, [IND] [f] will tell a story of the project; anticipate this.

— The story will reflect positively on all.

— The story will help us to intelligently defend against possible future PR attacks.

• If you want to write publicly about the project, sync with [    ] first.[g]

(3) Don’t spread private/sensitive/adversarial information; let the relevant people determine share permissions

• Don’t spread private information; use normal privacy practices[h]

— personal details

— psychological details

• Don’t spread sensitive information; let organizations decide

— legal information

— recruitment information

— financial/funder information

• Don’t spread adversarial information; let organizations and/or people who gathered the relevant information decide

— list of people to ask about before inviting, etc.[i]

— sociological case studies[j] - [    ]

— strategic bad actor information[k] - [    ]/[    ]

(4) Limit spread of CT, intention phenomena[l], and AI information

• CT — let [   ] decide

— It is fine to spread CT inside your own group

— Ask [   ]’s permission to spread beyond that (e.g., if you want to share [DOC[m]])

• Intention phenomena — exercise discretion

— It is fine to muse about intention phenomena inside your own group

— Exercise discretion in spreading it otherwise

— Whether to do bodywork, etc., is up to individuals

• AI — don’t help people build AGIs

— Don’t help people build AGIs

(5) Respect attribution

• Preserve attribution for major things (e.g., [SYS[n]])

• Don’t become rich and/or famous on someone else’s thing without attribution

(6) Document privacy

• Don’t publish someone else’s documents

• Can show but don’t share 454 slack quotes, Confluence documents, and Google docs; keep these private

(7) Security

• If you are keeping other people’s sensitive or private information, continue to use old ecosystem protocols, including:

— OnePass

— two-factor authentication

— phone security

(8) Common use of information/ideas/techniques

• So long as (3), (4), (5), (6), and (7) are respected, people may use, talk about, share with others, build businesses from, etc., the information/ideas/techniques developed by themselves and others

(9) Enforcement[o]

• Informal; we’ll keep each other in the loop if someone violates the information agreement

(10) Expected lessening

• Expect to revisit this in 12 months[p]

• Expect that the overall need for share restrictions will diminish, and that as a result we will wind down share restrictions over time, while still maintaining protection of sensitive information and people’s privacy

• If anyone concludes in the future that stronger information management is required, they should make efforts to educate others themselves, and should expect that that might be covered by some future arrangement, not this one

(11) Relation to legal

• This is not a legal agreement
— It is not meant to supersede or replace existing legal agreements

— It is meant to be based on norms of good behavior in society

— The signature below is to confirm having read and pledged agreement

Name _____________________________________

Signature __________________________________

Date __________

[a]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This purpose should not override your own safety and mental health.

We came together to try to achieve something we all thought was important.

We would prefer that the project's descendent Leverage 2.0, given its new structure, purpose and staff, not be intentionally damaged without good reason. But telling your genuine stories, and the truth about your opinions (including opinions that hurt Leverage 2.0) is okay by us and not something we want to prevent.

[b]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

Problematic meaning spreading information in ways that are especially and predictably likely to cause harm (e.g., spreading information that would enable people to cause harm with technologies, sharing information in a way intended to harm innocents, teaching psychological techniques in ways that seem likely to be damaging to someone, etc.)

[c]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

We feel intellectual property of others should still be respected through not sharing others' intellectual property or documents and providing proper attribution.

[d]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

Meaning we hoped at the time that this document would not overly restrict sharing of information.

[e]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

Treat this as something like "assume one of the better stories that are consistent with the information you have" or "try to use charitable interpretations of the actions and motivations of ex-colleagues" and not "don't tell the truth if it hurts someone" or "never share negative experience or concerns"

[f]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This refers to the name of an individual being removed for their privacy. Throughout the rest of the document, unless otherwise stated, blank brackets refer to individual names that were moved.

[g]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This was an attempt to help us slow down and calm down when tempers were high. [ ] is confident that they would have approved Zoe's story. We think that it is appropriate to drop this point now, several years after dissolution. You might consider generally useful practices to substitute now, such as "sleep on it", and "ask a trusted friend", but this item is not considered by us to be appropriate or binding in any way any longer.

[h]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This refers to following normal social conventions and standard interpretations of legal rules.

We still believe that it is important to respect individuals and that it is not okay to share personal information without the subject's consent.

This obviously includes confidential information received in psychological research contexts which should not be shared without the person's permission.

[i]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This refers to "Ask First" lists we maintained of people it made sense to ask first before inviting them to an event either because

(a) we had credible evidence of concerning behaviours (e.g., people seriously talking about wanting to destroy the world, reports of sexual harassment)

or

(b) they were important to the organization (e.g. donors).

From our conversations with nearby communities, we believe that practices such as maintaining lists or keeping track of individuals for this or similar purposes is common.

We think people should be free to make their own decisions about who is safe to invite to their events and share information based on their own models and own evidence. But we do not think is appropriate to share information about individuals on such lists based on other people's models or other people's evidence without their permission, particularly since they may have updated their perspective, model or evidence and so sharing such information may amount to unfair and negative gossip.

[j]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This includes case studies on individuals and institutions which are alive or active today, including some individuals who are potentially dangerous (e.g., Putin).

We believe individuals should use their own judgement and develop their own research of this type. We ask that individuals please ask the original case study authors before sharing.

[k]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This refers to the possibility that there could be potential global bad actors who could use the tools of rationality or strategic thinking to cause harm.

For an example of this concern,  see this case study: https://www.cnas.org/publications/reports/aum-shinrikyo-insights-into-how-terrorists-develop-biological-and-chemical-weapons

We think people should be free to think for themselves and make their own decisions about threat models of this type. We ask that individuals please speak to the ex-colleagues who developed the relevant threat models before sharing them.

[l]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

This was included to limit potential harms from psychology and to help with attribution/intellectual property.

[m]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

Brackets containing "[DOC]" have been used to denote where the name of a document was removed.

We are unsure if these need to be redacted but we think that should be up to the individuals wrote the documents.

[n]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

Brackets containing "SYS" have been used to denote where the name of a self-improvement system was removed.

We are unsure if these need to be redacted but we think that should be up to the individuals who developed the systems.

[o]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

To clarify, the means by which we intended to enforce this agreement is informally by keeping each other in the loop on violations.

Our only intention for this part of this policy was simply that if anyone violated the policy that information would be shared amongst the individuals that signed it.

[p]Leverage Research & Paradigm's clarifications and perspective today (Mon, Oct 18, 2021):

We apologize for not doing this.

We hope this discussion goes some way to rectifying this and revisiting the arrangement together.