Motivation for asking: From now on, I intend to use my answer here to continuously document new ideas I come up with instead of having them logged privately in Google Docs. This is part of my goal of reducing the time between conceiving of an idea and sharing it (How much delay do you generally have between having a good new idea and sharing that idea publicly online?).

18 Reactions

Mentioned in

163A central directory for open research questions

31Help us find pain points in AI safety

New Answer

New Comment

4 Answers sorted by
Top

dominicroser

Mar 09, 2020

WHAT: A book like "Strangers Drowning", but focused on the "E" of EA rather than the "A" of EA.

WHY: narrative can be such a tremendous force in changing people's lives. It's often more powerful than argument (even for brainy people).

There's already a lot of world literature and newspaper stories on people who have been tremendously altruistic. There is much less literature about people who have been tremendously altruistic and -- this is key -- have been motivated by their altruism to care about effectiveness and listen to the evidence.

I'd love to have a book with biographies or stories that traces -- in narrative rather than argument -- people whose love for others has pushed them to care about effectiveness, care about evidence, and generally care about a results-oriented outlook that focuses on what 'really works at the end of the day'. (Note that the book should not generally be about people who care about effectiveness and evidence -- but only about people who have deliberately chosen to do so out of altruism (rather than, say, out nerdiness)).

Possible biographies could include: Florence Nightingale, Ignaz Semmelweis, Deng Xiaoping, figures from EA and utilitarianism, some theologians in the 2nd world war who pragmatically looked towards ending the killing (Bonhoeffer, Barth, etc?), etc. Not vouching for this list of examples at all -- it's more to give an idea.

By the way, creating such a book could be a project for EAs with a different skillset than the cliché EAs.

Mati_Roy

Mar 07, 2020

moving my answers in separate comments below this answer.

particularly useful feedback includes, but isn't limited to:

links to a similar project that was already done
connection with people interested in this project
analysis of the usefulness of the project

note: those are just ideas, they might not be a priority, or good at all

Mati_Roy5y7

The Bullshit Awards

Proposal: Give prizes to people spotting / blowing the whistle on papers bullshitting its readers, and explaining why.

Details: There could be a Bullshit Alert Prize for the one blowing the whistle, and a Bullshit Award for the one having done the bullshitting. This would be similar to the Darwin Awards in that you don't want to be the source of such an award.

Example: An analysis that could have won this is Why We Sleep — a tale of institutional failure.

Note: I'm not sure whether that's a good way to go about fixing... (read more)

yhoiseth

Great idea! This sounds like a lot of fun. I'm also unsure about the net benefit. We might want to keep it as unaffiliated as possible from other EA organizations in order to avoid any spillover damage.

Mati_Roy5y4

Belief Network

Last updated: 2020-03-30

Category: group rationality; signal boosting

Proposal: Track people's beliefs over time, and what information gave them the biggest update.

Details: It could be done at the same time than the EA survey every year. And/or it could be a website that people continuously update.

Motivation: The goals are

1) to track which information is the most valuable so that more people consume it, and

2) see how beliefs evolve (which might be evidence in itself about which beliefs are true; although, I think most, including myself, ... (read more)

Mati_Roy5y3

Promise Prediction

Proposal: Have a prediction market on what politicians will accomplish in their next mandate.

Why: That way, it will make it easier for people to know how likely each policies are to be implemented, and it will make it harder for politicians to bullshit everyone.

Related: This project would complement really well the Polimeter which tracks the promises made by politicians. They are now part of the Vox Pop Labs.

Note: I think I've seen this idea somewhere else, but I don't remember where.

Mati_Roy5y2

Shaking hands across the world

Category: Bringing powerful countries closer together

Idea: Handshake statue in Time Square and some equivalent place in China, where people can give each other a handshake across the world

Effectiveness: I don't know;doesn't seem effective, but also maybe such symbols are powerful and would bring the world closer to each other, hence increasing cooperation / reducing risk of wars

Source: Space Force TV show, s1e7 8:30

Mati_Roy5y2

Royalty free AI images

Created: early 2019 (or maybe before) | Originally shared on EA Work

Cause area: AI safety

Proposal: Make a collection of (royalty) free images representing the idea of AI / AI safety / AI x-risk / AI risk that aren't anthropomorphizing AI or otherwise misportraying AI (both by searching for existing images and by creating more). This could be used by the media, local AI (safety) groups, etc.

Details: I think this is less of a problem than it used to be, but still think this could be valuable. If you want funding for that, you coul... (read more)

Mati_Roy5y2

Philanthropy tax / Giving your 2 percents

Meta-proposal: Research what would be the consequences of implementing the proposal.

Proposal: Give the ability to citizens to decide where X% (say 2%) of their tax goes directly (it can be a charity or a government program)

Details: Of course, government can rebalance the rest of its budget in such a way that there's no counterfactual changes. But maybe it would still make a change. If not, then maybe the X% has to go to a charity. Or maybe the donations could be made for more specific governmental projects.

Reas... (read more)

Mati_Roy5y2

Group for collective actions

Status: done, see: https://www.facebook.com/groups/LWCoordination/

Proposal: have a group to experiment with coordinating on small projects that require coordination

Example: I just posted a proposal about improving the Cause Prioritization Wiki. If 100 major edits get committed, then everyone does the edits they committed. This is useful because a wiki only becomes interesting when there's a lot of editors, so this allows the platform to get bootstrapped, and avoids the chicken-egg problem.

Comments: There's a meta-threa... (read more)

Mati_Roy5y2

Quantified Doomsday Clock

Context:

Since the Doomsday Clock from the Bulletin of the Atomic Scientists doesn't have any clear methodology for why the clock advances or recedes, I am providing the Metaculus Doomsday Clock as an alternative. Currently the way it advances is by using the Metaculus median prediction of humanity going extinct by 2100 to determine how many minutes we are from midnight. It can be improved, so make suggestions in the comments.

https://sites.google.com/view/metaculus-doomsday-clock

(source: Matthew Barnett's Facebook wall)

P... (read more)

Kirsten

Rather than 2100 can I suggest the next century? Otherwise we'd move away from midnight as we move toward 2100 - very counterintuitive

Mati_Roy

yeah good point, I agree; thanks!

Mati_Roy

https://aicountdown.com/ links to

Mati_Roy5y2

Impact of the 5% payout rule

Category: meta-EA; research

Proposal: Research what would be the consequences of removing the 5% payout rule.

Motivating intuition: maybe it would help longer-termist causes (?) and it might also increase the global ratio of investing / consumption (?)

Date posted: 2020-03-06

Additional information:

A foundation must pay out 5% of its assets each year while a public charity may not.

Donors to a public charity receive greater tax benefits than donors to a foundation.

A public charity must collect at least 10% of its annual expenses fro

... (read more)

Mati_Roy3y1

Increase the prize for the International Mathematics Olympiads

Rationale: It's a useful source of talent EAs have used, and the current prizes are pretty low (less than 100 USD each AFAIK).

I'd be willing to pitch in that prize. Please reach out to me if interested.

Mati_Roy5y1

Rationalist Olympiads

Potential funding: EA Meta Fund

Ideas:

See section “Games and Exercises” of “How to Run a Successful LessWrong Meetup Group” for some ideas: https://wiki.lesswrong.com/mediawiki/images/c/ca/How_to_Run_a_Successful_Less_Wrong_Meetup_Group.pdf
Calibration game
Cognitive science quiz (ex.: game Irrationality)
Cognitive bias game (note: I’m still developing it)

Mati_Roy5y1

FDA Policy Think-tank (and/or advocacy group)

Mati_Roy5y1

Science policy think tank (or advocacy group?)

Potential problem: it might accelerate all scientific progress, which isn’t relevant in the framework of technological differential progress, or possibly harmful (?) if, for example, AI parrallelizes better than AI safety

Mati_Roy5y1

Sober September

Created: early 2019 (or maybe before) | Originally shared on EA Work

Cause area: aging

Dry Feb is a Canadian initiative that invites people to go sober for February to raise money for the Canadian Cancer Society: https://www.dryfeb.ca/.

Imagine this idea, but worldwide and for general medical research.

I would suggest fundraising for the Methuselah Foundation for its broad approach. They fund a lot of prizes which create market pressures for medical progress, so avoid the donors to have to figure out which research groups are the most effective.... (read more)

Mati_Roy5y1

Decision Theory Interactive Guide

Created: early 2019 (or maybe before) | Originally shared on EA Work

Proposal: I think this could help understanding decision theories (especially functional decision theory). There could be some scenarios where the user has to choose an action or a decision procedure and see how this affects other parts of the scenario that are logically connected to the agent the user controls. For example: playing the prisoner dilemma with a copy of oneself, Newcomb’s problem, etc. Could be done in a similar way to Nicky Case's games.

Mati_Roy5y1

EA StackExchange

Created: early 2019 (or maybe before) | Originally shared on EA Work

Create a quality StackExchange site so that the EA community can build up knowledge online.

Note: The previous attempt to do so failed (see: https://area51.stackexchange.com/proposals/97583/effective-altruism).

Mati_Roy5y1

Maybe summarizing the book "Who Goes First? The Story of Self-experimentation in Medicine". Two possibly important thesis:

self-experimentation is important
medical innovations are available way before they get adopted

Mati_Roy5y1

Category: research

Externalities of war predictions

See: link

Mati_Roy5y1

Moved from my short form; created on 2020-02-28

Group to discuss information hazard

Context: Sometimes I come up with ideas that are very likely information hazard, and I don't share them. Most of the time I come up with ideas that are very likely not information hazard.

Problem: But also, sometimes, I come up with ideas that are in-between, or that I can't tell whether I should share them are not.

Solution hypothesis: I propose creating a group with which one can share such ideas to get external feedback on them and/or about whether they should be ... (read more)

Kirsten

Why wouldn't you just ask four people who you trust to review each idea in confidence? Why formalize it or insist they reciprocate it?

Mati_Roy5y1

Altruist credits

Epistemic status: not sure if the idea works

Category: meta

Proposal: Pay someone with a 'donation gift card' or 'donation credits'

Details and rationale:

Often, when I work on a project approved by EAs, I don't necessarily want to be paid as much as I want to be able to have people work on my EA projects in the future.

Imagine you have a Donor Advisor Fund called the Altruist Bank which emits one Altruist Credit per USD you put into it. The Altruist Credit can be spent by saying to which charity you want the DAF to sen... (read more)

Mati_Roy5y1

Coronavirus: Should I go to work?

UPDATE: An EA project I'm part of might do this

summary: have an app that helps people decide whether they shouldn't go to work

context: in the last 12 hours I spent maybe about 2 hours 'empowering' someone I know by giving them more information to help them decide whether they should take sick days

problem: knowing what's the probability one's infected (by the coronavirus) helps informing them about whether they should avoid going to work. the probability beyond which you should stay home is not ... (read more)

Mati_Roy5y1

Forum Facebook page

Posted: 2020-03-07

Category: signal boosting

Proposal: Share the best (say >=100 karmas) posts on the EA Forum on a Facebook page called "Best of the EA Forum"

Why? So that people that naturally go on Facebook but not on the EA Forum can be exposed to that content

Note: If there's a way to get this list easily, it might facilitate the process.

Update: 2020-04-24

Experimental page using Zapier: https://www.facebook.com/EAForumKarma100/

x-post: https://www.facebook.com/groups/1392613437498240/permalink/2947443972015171/

Aaron Gertler 🔸5y6

I appreciate this idea! However, I'd prefer that people cross-post Forum content to groups that already have substantial/relevantly targeted audiences (actually, I'd really like people to do this more often), rather than creating a new group that could split off some of the Forum's readership.

Having a Forum-focused Facebook group also seems like it would raise the chances of more discussion happening on Facebook rather than on the Forum posts themselves, which seems bad (comments harder to find later, not linked to anyone's profile, not open for karma voting, not eligible for the Comment Prize, etc.)

If the group really is just a collection of links that people can easily share in other groups, and if discussion comes back to the Forum, it could be a net positive. I'll be curious to see how it gets used.

Mati_Roy5y2

thanks for your comment, I totally agree!

maybe we could ban comments? and delete the page if that doesn't end up working?

Aaron Gertler 🔸5y3

Rather than a ban, probably official discouragement + a polite reminder that people should add their comments to the Forum as well as the Facebook posts? If people really want to talk on Facebook, it seems bad to stop them, but gentle nudges go a long way!

Mati_Roy