This post describes an “EA common application” and some perspectives about it. 

This post has been written quickly, to encourage and support any current or upcoming efforts to build a very high quality common application for Effective Altruism. 

I am not involved and I am not aware of any team currently working on this.

1. Introduction

EAs should be aware that there is an opportunity to build a “common application”, an extremely important and impactful institution for EA. 

A common application is an institution that coordinates EA hiring.

This institution could be immensely valuable. One of the reasons it could work well is that only a fraction of qualified talent is hired in any recruiting process. But EA values talent more broadly and valuable candidates should be developed and supported beyond any one hiring cycle.

The best versions of the EA common application go beyond streamlining the hiring process, and can greatly increase the number of talented and engaged EAs, and create better matches and opportunities for EAs as the movement grows. 

As a servant to EA organizations and funders, the common application also gives a way to increase vetting and coordination capacity and improve the happiness and confidence of EAs, supporting the effectiveness and health of EA as a movement.

A version of the common application was almost funded by a major EA grantmaker this spring, but the founder left the project for reasons specific to themselves. This funding included near-six figure salaries for a staff of three: architect, developer and coordinator. Even more importantly, this EA grantmaker offered the possibility of integrating their talent pipeline, one of the best talent pools in the world. The integration of this EA talent pool would be crucial to creating an extremely high quality implementation of a common application.

2. Tradeoffs against showing grant making process (meta comment)

This is one of three documents related to the common application. This post you are reading was written for the EA forum. The other two documents are “live documents”, documents actually used in the grant making process. One is an exploratory document, designed to maximize input, feedback or criticism, and the other one is the grant proposal itself.

These two “live documents” can publicly illustrate one potential process to getting a successful proposal, grant, and EA institution created^[1]. 

Ideally, I wanted to use these documents to show the grant making process. But there is a tradeoff between using these to illustrate the grant making process and encouraging the creation of the EA common application^[2]. I think there are important introductions to be made about EA funding, or maybe misconceptions to correct about grant writing or grant making. Carefully situating these documents into a narrative seems valuable. But this is hard to do well in a short period of time.

I feel responsible to get information about the common application out. Also, these documents are currently being shared in an ad hoc way, and it seems better to make them public.

This post you are reading might be edited and changed heavily.

3. Disclaimer about writing and request for feedback

This document is quickly written, is incomplete in many ways, and might be hard to read. I don’t think there is any template for building a complex institution and I can’t immediately think of anyone who has written similar advice.

I think the audience of this document are EAs who are or might become highly engaged in building the common application, and I hope it should still serve those people. 

This disclaimer isn’t an excuse or trying to avoid criticism–it’s exactly the opposite. If anything seems confusing or if you have any questions at all, please comment!

The rest of this post contains two major sections: one quick overview of the common application, and then a number of considerations for one vision of the common application.

4. One presentation of the EA common application

(Note that the following describes one vision of the common application, and is dependent on founding team preferences and ability. The actual implementation will be different, even if the team executes the project perfectly.)

Basically, the “common application” is a common point of entry for EAs and talented individuals applying to EA organizations. 

To be as concrete as possible, in the first year or so, a common application might include two things:

• A website that is used by applicants and EA organizations
• A team that is widely seen as competent, principled and transparent and has engagement from major EA organizations

To give a down to earth description, this would be a website that everyone uses and applies to, when working in EA. It is seen as the obvious, optimal thing to do. Everyone is happy using it, including talented candidates, as well as sophisticated leaders of EA orgs, of every background and EA purpose.

To these organizations and applicants that are users, the site will be simple and straightforward.

But for the founders/creators, achieving strong implementations is harder than it sounds, and in the best version, there are (extraordinarily) complex considerations^[3]. 

Good implementations of the common application are demanding, but equally valuable to EA. For the founding team, there is an enormous opportunity for skill expression and a high skill ceiling for technical ability and institutional knowledge.

4.1 Operational focus is crucial

The bread and butter of the common application, which will dominate the work of the founding team in the first year or more, is the day-to-day work to get operations running smoothly, and to build expertise and trust among organizations and candidates. 

This is important for the success of the common application in later stages. Also, while it seems mundane, just having experienced, trusted staff have friendly contact with talented candidates is a valuable role of the common application. It is essential to have founder(s) who respects and will execute this work well.

4.2 Essential Work (year 1-2 goals)

In the form it will take in the first year or two, the value of the common application to EA includes:

• A streamlined, common website used by thousands of talented people looking to contribute or work at EA orgs, as well as a competent institution that provides services, advice and standards to EA organizations.
• A central place that provides insights about EA recruiting (similar to this post, but for all EA orgs, all the time), and observes and can react to bad outcomes ("really hard to find an EA job"). Later on, the common application can react to gaps as well as surpluses for talent at the movement level.

4.3 Further Outcomes (activity in years 2+)

Given good execution of the above goals, with skill and a strong team, the common application can provide enormous and unique value in later stages:

• Working as a servant to EA organizations, the common application can do the delicate, difficult task of developing assessment, screening and guidance tools. To be concrete, it could administer friendly assessments that gently provide feedback and advice to a newcomer to EA. These tools for candidates and organizations would make EA organizations recruit more effectively and provide confidence and insights for EA talent.
• The common application can go far beyond streamlining recruiting by bringing strong candidates into EA, and making better matches for existing talent, for example, by creating new roles, catching candidates who might bounce off EA and building up deep pools of talent that endures beyond any single job search.
• With careful design and working with existing grantmakers, the activity in the common application will provides a natural, powerful way to further develop and grow the pool of EA “vetting”, trust and communication that is important for EA scaling, supporting existing strong EA culture, norms and institutions

This vision of the common application is unusual. It’s hard to think of any other movement that has an institution like this, or that could build one. In later stages, some of the ideas, methods and practices developed could be historic and groundbreaking. 

The previous writer/founder had interest from professors in Stanford GSB , Sloan/MIT and Penn State, as well as other schools, who expressed interest in working for free, studying and developing methods (mechanism/market design, matchmaking/assessment) for this common application (because the work and data can provide many publications).

A common application builds on some of the best traits of Effective Altruism: impartiality, the consideration of others and their contributions outside of one's own organization, the high level of coordination and communication between EAs and organizations, and a valuable talent pool that should only increase in value.

EA’s funding situation, and future expected growth, means that the current time might be an ideal or unique opportunity to create the common application.

A strong common application provides an enduring asset for the movement, a pillar that provides stability, confidence and happiness, and enables object level work for hundreds or thousands of people in the coming years.

5. Final disclaimer about too much writing 

I am uncertain, but it seems good to be cautious about creating the impression that writing a lot, or being publicly articulate about the process of setting up an organization, should be valued. 

Beyond a certain level, great effort by founders on writing, or excessive reward to writing by the community, might hamper the creation of high quality organizations in some way^[4].

I am writing this because the following content tries to help, but there is a danger of being prescriptive or setting the wrong tone, in some way that I don’t understand.

6. General thoughts

6.1 Brief thoughts on founding team and initial activities

I think it seems reasonable that the founding team for the common application should have the following traits:

• Software development experience (while rare, the ideal would be past development of an user facing app from scratch)
• The founding team should contain longtime EAs, who ideally are known by the major EA funders
• The founding team should have experience with, or are known to many EA organizations among every major EA cause area

These above traits are valuable, but I’m not sure all of these are absolutely necessary. 

Working backwards, based on some of the activities of the team in the first year, the further list of traits or behaviors seems good:

• Someone on the founding team needs to have an outgoing personality, and strategic and soft skills, such as excellent listening abilities, so they can build trust and consensus among a diverse group of EA organizations.
• The team needs to execute operations diligently, and plan carefully for future work and functionality.
• For at least a few key topics, there should be deep insight and careful strategic planning. Some decisions will be made that affect activities a long time away^[5].

Note that the founding team does not need to have all this talent before funding is granted. They can add or hire afterwards. The founding team could be just one person.

Later documents in this series, such as the grant proposal, specify the founding team, giving the titles and roles explicitly for three people. This detail will be shown (but if you want to know now, ask in the comments). I’m unsure if specifying this level of detail is good, so I haven’t listed it out here.

6.2 It’s probably better to wait to design and implement novel assessment ( new quizzes, work trials, etc.)

One feature of a common application that many people think of immediately is candidate assessment. For example, people think of creating work trials or “TripleByte” style quizzes that measure aptitudes, and guide candidates to jobs.

However, I think assessment is deceptively hard to implement (this was a major issue raised by others in getting feedback). 

If you look at actual implementations of assessments, you can get a sense of concerns: the tests in some disciplines are often disliked (Leetcode) and have seen setbacks (SAT). At Triplebyte itself, there seems to have been a pivot, deemphasizing assessment (this detailed article gives a sense of the considerations). When giving feedback, one experienced EA raised concerns, referencing drawbacks with Chinese imperial examinations.

I think the challenges and downsides of assessment are large and include unintentional coloring of the EA talent pool, creation of myths, like certain background or skills are preferred in EA, or that certain levels of ability are needed. These issues require  care and caution.

6.3 Notes on original funding “offered” and founder skill

To support the founding team, this section is more candid and detailed than is preferred.

Some details about funding:

• The funding approved was for three people with nearly six figure salaries, and extra funding was provided for appropriate expenses (legal, contracting for specialized services, etc.). I think salary level could be increased or decreased with justification, and was not the limiting factor.
• I would be skeptical of a much larger or much smaller founding team (just 1 person or 5 or more people seems bad).
• Having an impressive team (track record, positive sentiment from EAs of good judgment, and strong outside options) is probably important for funding and approval.

Subthread about boosting salary for a developer

The writer/"founder" didn’t claim to be "technical". Understanding that the current market of developer is tight, they mentioned the possibility that extra funding was needed to increase the salary of a developer they would need to hire. This would raise that salary well above that of the other founding members.

This large salary increase for a developer on a founding team has several considerations:

• I think the primary driver of hiring a developer (or any "early hire") should be "founder skill" and attractiveness of their vision and leadership to discerning talent. This ability is the key traits of the founder/founding team.
• In the most principled, benevolent sense, I think a strong founder should be able to bend down salaries and attract strong developers, especially given the potential of this particular project. In my view, for a less established project that should offer a lot of independence and room for growth, it seems bad to use EA money to buy expensive talent at market rates.
• At the same time, the current market for strong, experienced developers is tight, and justifies a large salary.

In the end, I expect the "founder" would have been able to increase the salary to near market rates for a good developer. 

All of the above considerations in this subthread apply for a "nontechnical" founder. A founding team with one or more developers wouldn’t have this problem (with the tradeoff that other skills in the founder might not be available).

6.4 This is a major software project

Other people, describing some other visions of the common application, seem to describe a fixed and limited amount of development hours. 

That’s really different from this vision of this common application and I don’t think it would work.

To see this, the CEA and LessWrong team currently has multiple developers, and a budget into the 7 figures (with many staff not at market rates). This investment and spending is extremely valuable and includes multiple strong technical leads, various developers and designers. This is all based on a technology, UX, and other patterns that seem established (Reddit, internet forums).

Most versions of the common application will probably require strong SWE involvement. The initial, founding team should include at least 1.0 FTE software developer, with broad experience.

6.5 Software engineering support and forward looking culture development

This describes some actions the non-technical writer/"founder" of the original team may have tried to take to support their developer hire(s):

• They may have lined up multiple senior FAANG developers as mentors and advisors ("Senior" or "Staff" in their title, manager with hiring and software architecture experience)
• They may have lined up several strong founders of small companies, who had good exits. These people were undergraduate CS students, graduated and solo developed projects, often bootstrapping, into good exits (maybe 7+ figures)^[6].

These two types of mentors or advisors should complement each other. One has more recognition and experience with politics, stakeholders and maintenance of systems in an enterprise. The other type is scrappy, has experiences with customers and end user features, and owns multiple systems, including operations of their entire company, from founding to exit, while working in small teams.

These people would serve as mentors to the developers on the founding team and advisors to the project. 

I think this would help attract and retain valuable developer talent. It would also help ensure health and success of the software development team that is extremely important as the common application scales.

It still seems valuable for scaling, but it’s less clear this is important if the founding team contains experienced SWE.

6.6 Use of economists and business school faculty

One asset to the project would be advice or involvement from economists or business school professors with academic and practical experience in subdomains like “market design”,  “mechanism design” or “matchmaking”^[7]. 

This talent is formidable. For expert faculty in these subdomains, many have worked in developing actual systems in different institutions and diverse cultures in places, such as the US, South America and Israel. 

These people have gritty, practical experience: in designing a Boston school choice system, one professor mentioned how a small UX misconfiguration, inappropriately showing one option too early, ended up risking the outcome of the entire matchmaking system. 

I think that many EAs will be able to intuit, design and solve most problems and create systems without any help. 

Even so, for pragmatic reasons, this outside professional talent is probably valuable to the founding team of the common application. 

The claims of value in the common application are large, tend to be abstract, and may be impractical to communicate quickly. The credentials of the professors and the fact they are interested and working for free^[8], is a strong signal that these outcomes are achievable. This probably helps relations with many stakeholders downstream.

6.7 Improving and scaling trust and vetting in EA

I am very uncertain about this section, but it is good to think ahead about the ways the common application can improve trust and vetting in EA, which are limited and constrained:

• One way the common application can improve vetting is the direct utility of assessment tools that could be eventually developed, such as tests, interviews.
• Over time, after developing trust and strong relationships with many EA organizations, the common application should have access to ways of measuring talent (usually through EA organizations themselves). For example, maybe at some point, strong lawyers, public relations experts, or mathematicians may need to be hired by EA organizations. Once the patterns to assess them are developed, this ability can be shared and used to advise and assess promising new ideas or projects. This is particularly valuable because this ability should emerge naturally, through the object level work of successful recruiting of extremely high quality talent.
• Finally, while many EA organizations used blinded screening for many applications, some processes may require and involve explicit vetting or recommendations. This may create a role for a pool of vetters that assess and guide talent into EA. This pool can be developed deliberately to develop vetting more broadly for EA (for example, people can rotate through the pool to gain experience). Note that this pool and vetting can be used support, not “gatekeep” EAs from opportunities—to see this one way, existing EAs might have many friends to recommend them, but many valuable newer EAs may not have these connections. A vetting pool, experienced and trusted by EAs, substitutes for these networks and aids the input of valuable talent into EA.

I am very uncertain of the above. I think the actual implementations, the downsides of these ideas, and opportunities that might emerge are hard to predict. 

However, I strongly believe there is an opportunity to expand vetting and trust networks through the careful creation of the common application, and these ideas give some sense.

7. Questions?

Do you have any questions?

• Do you have comments or criticisms of the common application itself?
• Was any content in this post confusing or unclear in any way?
• Do you have any other questions or topics you want to see discussed?

I hope the writing above should make sense. 

If you have any questions, please ask!

Some of the writing of this document has been aided by Written.ai^[9].

1. ^{^}

   I think this shows one of many different styles and processes, and the actual process for getting funding can be very different. This can be much longer or shorter for other people.

2. ^{^}

   For example, in complicated ways, some content in the live documents was to address topics that are idiosyncratic to the writer/"founder", or the project itself. This content could be net negative in another grant proposal and it would be good to edit or flag it. There might also be tradeoffs in censoring sensitive content in the “live documents” to improve confidentiality or prevent over indexing.

3. ^{^}

   I think one issue is that there are tension between functions of the common application: 

   • One of the key abilities of the common application is sharing applicant interest and progress among organizations, e.g. there might be 10 extremely talented candidates who got rejected in the last stage of a hiring round. This talent can be retained in EA and hired in other organizations, or even encouraged to start their own projects.
   • At the same time, while sharing all of this information, the common application should not disadvantage applicants. Applicants' general comfort and happiness and the instrumental value of their information are both crucial. This is hard to achieve because adverse information can leak in complex, implicit ways
4. ^{^}

   Note that this disclaimer tries to flag the topic of giving status or approval of organization creation based on ornate, public writing. However, communication is important in organizations. For example, honest public or community relations, customer service, addressing stakeholders privately, and “front womaning” by the “CEO”, “lead” or “director” of an organization, all seem crucial.

5. ^{^}

   Examples of decisions might include when to choose to pursue development of talent assessment, an important topic discussed elsewhere in the document. Other long term decisions might include how and when to develop norms for information sharing by candidates; which organizations in particular to initially focus on serving or partnering with, or not having a focus on any organization at all.

6. ^{^}

   These particular founders have a quality of being worldly yet kind, and have a “servant mentality” and value personal coding. These founders are Canadian. Culture, capital and exits in Canada are different (Silicon Valley money is many times bigger).

7. ^{^}

   Note that other business disciplines can have strong and relevant experience. One discipline is “marketing”, which has a strong applied math subfield with scholars who study hiring and matching of talent in the real world. 

8. ^{^}

   The previous writer/”founder” had interest from professors in Stanford GSB , Sloan/MIT and Penn State. 

9. ^{^}

    Written.ai claims to be a tool that uses NLP and other tools to help thinking and writing.  In a recent conference, ten AI safety researchers of various associations and beliefs were given a description of Written.ai. No one said the development or activities described seemed dangerous from an AI safety perspective.