Hide table of contents

My team (Rethink Priorities’ General Longtermism Team) is aiming to incubate 2-3 longtermist projects in 2023. I’m currently collecting a longlist of project ideas, which we’ll then research and evaluate, with the aim of kicking off the strongest projects (either via an internal pilot or collaboration with an external founder). 

I’m interested in ideas for entrepreneurial or infrastructure projects (i.e., not research projects, though a project could be something like “create a new research institute focused on X”). 

Some examples to give a sense of the type of ideas we’re interested in (without necessarily claiming that these specific ideas are particularly strong): An organization that lobbies for governments to install far UVC lights in government buildings; a third-party whistleblowing entity taking reports from leading AI labs; or a remote research institute for independent researchers. You can see a list of our existing ideas here.

I’ll begin reviewing the ideas on April 17, so ideas posted before then would be most helpful.

New Answer
New Comment

4 Answers sorted by

I see that the compilation and distribution of "civilisational reboot manuals" is already on the list. I love the concept, but think this scope should be significantly expanded to include stress testing and refinement of the drafted content. This would verify whether the most important facets of knowledge and technology are covered, and if the detail and style are such that they can be followed. I heard this suggested by Lewis Dartnell (author of "The Knowledge") on the 80,000hrs podcast, and think it would be great to really run with it. A fun, high-profile and potentially profitable way would be through a televised competition format, where teams of "survivors" have to try rebuild as much of the tech tree as possible (or reach a set technological achievement), with a "civilisational reboot manual" as their guide.

The mechanics of such a competition would need thoughtful planning to get a working balance between being sufficiently realistic of civilisational collapse scenarios (number of people, resources on hand etc.), have an acceleration mechanism to model decades of rebuilding within a season length, and be watchable. Challenging, but I don't think it would be a show-stopper (terrible pun, sorry). 

Benefits of this could include: 

  • Raising awareness of various existential risks. Perhaps each season/team could model a different collapse scenario such as nuclear winter, engineered pandemics, AI misalignment and so on, with the opening sequences explaining the likelihood of these events occurring and what action needs to be done to prevent them. I acknowledge that broadcasting to a potentially global audience could be a reputational risk to EA, and would have to be managed carefully.
  • Stress testing many of the assumptions we have around collapse and rebuilding scenarios.
  • An opportunity to get funding and visibility for larger scale testing of proposed technologies and solutions. E.g. for a "nuclear winter" scenario, trialing some of ALFED's research on simple greenhouse construction, seaweed/mushroom farming and the like.
  • Learn where the gaps are proposed civilisational reboot manuals, as there are likely some we cannot anticipate until they are tested realistically. In my head I see someone trying to recreate a particular machine or chemical process, but one small component isn't described in sufficient detail and everything grinds to a halt. 
  • Study what skill sets are needed amongst "survivors" and what governance structures work well, to ensure both progress and relative harmony. Some of these may be counter-intuitive.
  • Experiment whether the provision of select tools/technologies dramatically accelerate the tech tree rebuild. For example, a good blacksmith's anvil is very difficult to make from scratch, but once you have one it lasts nearly forever and facilitates the creation of innumerable useful items. These could be then be included in "reboot kits" along with the manuals themselves.

An incubator team could refine the concept and goals, perhaps do some limited trials, and then pitch to various networks or streaming services. 

 

Physical engineering lab to build capacity for prototyping hardware ideas with relevance to areas identified as important to the long-term future

https://forum.effectivealtruism.org/posts/9BDzFqAXu7sqPvRn5/reslab-request-for-information-ea-hardware-projects-1

An organization that identifies cities/regions across the world that are in danger of constructing a new power plant in the next 5-10 years and lobbies for the construction of a virtual power plant instead. 

In terms of wild animal welfare, I felt frighted when I've first seen the extent of using barbed fence in Central America (and probably American continent in general, I assume). There were published several field studies mainly from Australia, in which authors tried to evaluate harmfulness of barbed fence on wild animal health and results were relatively sound. It came especially bad for big mammals such as kangaroos, dogs etc. and winged mammals such as bats and flying foxes. Personaly, I've seen couple dogs in Costa Rica with damaged eyes and cut injuries on the body. The barbed fence could had been the cause.

Because extend of using barbed fence in Central America is vast and apparently preference for using it is deeply rooted into the culture of locals (on the rural and even small cities majority of people use barbed fence as a common way to demarcate their property), eradication of this custom and agency for replacement for more compasionate version of fence to wildlife, will be a longterm project. It almost certainly won't happen in a few years. And ultimately I would love to see the change in policy of using barbed fence worldwide, not just in the Americas.

[comment deleted]10mo1
0
0
Comments2
Sorted by Click to highlight new comments since: Today at 9:31 PM

Wonderful work! I’ve commented directly on the spreadsheet, but for the benefit of anyone who won’t check it:

Several of these ideas can be rolled into one:

  1. A remote research institute for independent researchers
  2. Infrastructure to support independent researchers
  3. Building vibrant EA academic communities in Africa, Asia, and Latin America
  4. AI alignment prizes, advance market commitments, and other forms of proto–impact markets

The scheme that I imagine could have all of these benefits:

  1. Find and recruit more independent researchers for high-impact research
  2. Tap into talent pools in countries that don’t have a lot of EA presence
  3. Circumvent restrictions on foreign donations/grants to researchers in some countries
  4. Support independent researchers monetarily
  5. Kickstart academic careers
  6. Support many small research projects efficiently (low due diligence overhead)
  7. Recruit for-profit investors such as business angels and impact investors to derisk research for researchers
  8. Derisk research for potential risk-averse non-EA funders
  9. Help researchers network, find potential advisors or collaborators
  10. Provide researchers with infrastructure (servers, labs, etc.) efficiently
  11. Monitor and improve the counterfactual impact of prize contests
  12. Tap into corporate funding for prize contests (in high-growth industries) 

The best thing is that to my knowledge it should be fully legal to do this.

We (GoodX) are working on infrastructure to support independent researchers with funding and simplify grant applications. We’re not currently implementing this particular scheme, but that could change given the right team (experts in US securities law and startup fundraising).

The approach:

  1. Build a network:
    1. Set up a nonprofit think tank with some form of limited liability and a suitable purpose so that it is exempt from the requirement to register any publicly traded securities with the SEC.
    2. Network among business angels, HNWI (including non-altruists), possibly VCs. Even $100k go a long way in low-income countries, so “high net worth” can be a low bar depending on the country.
    3. Watch out for prize contests, AMCs, governmental and private outcome payers, etc.
  2. Let the investors scout out great researchers:
    1. An investor could be an Indian economics professor – smaller high-context investors can lead; larger low-context investors can follow.
    2. Possibly help with the match-making, especially once we have a mature network ourselves.
  3. Match-make between investor + researcher and prize contest/AMC/outcome payer:
    1. Make contracts with large funders such as Open Phil, Gates Foundation, USAID, etc. over outcome purchases, AMCs, etc. that match the areas of expertise of the researchers.
    2. This could even work for risk-averse funders who could not otherwise support scientific research.
    3. Or enter the research into existing prize contests.
  4. Take the investment from the investor, pay the researcher a monthly contractor salary, hold some money back to cover costs.
  5. If the researcher is successful and the outcome payment is disbursed, it goes to the investors.

Meanwhile nothing keeps that think tank from also seeking grant funding and using its network to pay contract researchers from that. Especially researchers who have proven themselves in a prize contest but can’t currently find any new suitable outcome payor, could be kept under contract from grant money.

We’re always happy to have calls on such topics!

I appreciate this initiative, @Buhl .  I went to the google form and noticed it requires permissions to update. There are a lot of entries on it, and it looks like the last update was January 2022, pre FTX crash. Not sure if others felt this way, but personally this made me question whether reading through the existing ideas /coming up with new ones would be a good use of time versus parasocial.

Is your goal to find out which ideas have the greatest support on the forum, to generate more ideas, to find people interested in working on particular ideas, or something different? 

Are you hoping people will compile and upvote/ downvote ideas here in the comments?