Fellowship alternative: idea synthesis scenarios
Introduction
In response to this post, I am suggesting a fellowship format where participants review different materials and discuss complementary (sub-sub-)questions, which synthesize into broader queries. This leverages the benefits of specialization while allowing re-specialization anytime one participates. An idea synthesis scenario that covers hundreds of sources and questions can be done in a day, which makes it ideal for accelerated retreats.
I am including an example with 216 questions and materials, which summarize to 36 sub-tasks, 6 practice tasks, and 1 broad question (about opportunities in EA). At most, 216 people can participate but at least 36 for maximal efficiency.
I would most appreciate feedback on this idea as well as on the material and question choice in the example.
Questions, tasks, and materials
1 Question
What are some opportunities in EA?
6 Practice tasks
1: You received $1 million to check whether EA is not missing any crucial considerations. How are you going to invest these to maximize the expected value?
2: Draft an introductory speech to Our Common Agenda UN discussion.
3: Draft a memo for the hypothetical EA University of Liberia in which you suggest several organizational models, recommend epistemic upkeep practices, outline risk mitigation frameworks in EA, and offer a way to develop absolute advantage.
4: The International Organization for Migration Indonesia is overwhelmed by the increasing (it is 2030 and climate measures have not significantly progressed over the past decade) climate migration. Decide on whether and if so how EA should improve or prevent the situation.
5: There is a non-EA funder with $100 million interested in malaria reduction in the Democratic Republic of the Congo and generally global health and development. Develop talking points for an expert interview call they scheduled with you that motivate the funder’s optimal investment and further research.
6: The new rapidly scaling cricket farming in the Amazon company is selling dividend stocks. How much are you going to invest and (if anything), how are you going to invest your dividends?
36 Sub-tasks
1.1: Define EA fundamentals which should not change with any new findings.
1.2: Discuss the progress on crucial considerations in EA.
1.3: Discuss what resources should be focused on reducing the expected harm of a nuclear exchange.
1.4: Outline discount rate calculations which may appeal to various groups.
1.5: Discuss which entities should be given legal protections and liabilities.
1.6: Discuss the cost-effectiveness of lab meat investments.
2.1: Draft the 2030 update to the Sustainable Development Goals.
2.2: Describe how EA goes about developing key partnerships.
2.3: Discuss the benefits and costs of different networking strategies with the Impact Hub.
2.4: Describe how experts contribute to the development of AI regulatory framework of the United States.
2.5: Briefly describe the conflict in Ambazonia and viable ways forward.
2.6: Discuss how countries use human and artifial intelligence to safeguard peace.
3.1: Discuss AI alignment particulars.
3.2: Sketch and evaluate community organization models.
3.3: Describe current epistemic norms in EA and draft aspirational ones.
3.4: Summarize some discussions on risk in EA and suggest an optimal risk mitigation strategy.
3.5: Suggest biosecurity investment risk framework.
3.6: Draft EA diversity metrics.
4.1: Brainstorm catastrophic risks beyond those popular in EA.
4.2: Describe the means and objectives of AI persuasion.
4.3: Brainstorm new systemic change NGO ideas.
4.4: Describe thinking about long-term impact metrics in EA.
4.5: Brainstorm animal welfare courses possibly popular in China.
4.6: Find ways to prevent or institutionalize spending large sums on refugee survival.
5.1: Rank complex emergency programs by their cost-effectiveness.
5.2: Order marginal spending on neglected tropical diseases by its cost-effectiveness.
5.3: Rank malaria solutions by their cost-effectiveness (consider long-term impacts).
5.4: Arrange EA-related organizations on the scale of Perpetuate suffering the most to Reduce suffering the most.
5.5: Estimate the position of several UN agencies on a scale from ‘Perpetuates discrimination’ to ‘Eliminates discrimination.’
5.6: Draw funding and direct and indirect benefit flows of major development banks on a map.
6.1: Brainstorm ways to understand insect neutral point or a more relevant concept explaining their consciousness states.
6.2: Brainstorm best cricket farming practices.
6.3: Brainstorm sufficiently high wellbeing non-human animal farming metric values.
6.4: Brainstorm cost-effective Brazilian rainforest interventions (consider wild animal welfare impacts).
6.5: Brainstorm ways of increasing wild animal welfare without reducing birth rate.
6.6: Brainstorm climate change mitigation programs that may be popular in Brazil.
216 Questions and materials
See the table here for the questions and materials. Column A includes 36 actual tasks, special tasks, and functions for a scenario feel.
This sounds like a really interesting idea! I’m ambivalent about ‘reading group’-style fellowships, so I’m glad people are coming up with other ideas. I’d love to hear more about how you envisage this working in practice.
Thanks!
One fun way could be maybe 2-hour block at events (4*30 mins). Virtual events are another option (more affordable than in-person, can take longer because attention is not so scarce). For a more chilled atmosphere, maybe a weekend retreat where a third of the program is this scenario and ideally multiple groups run it at the same time, to combine insights and meet more friends.