A way to implement efficient Futarchy for small groups, on par with how easy it is to run small groups via democratic vote.
Like voting, but with more explicit rules for good deliberation and delegation of deliberation.
Rather than voting, deliberate until unanimous consensus is reached, forcing the group to address minority concerns.
Add to the previous the idea of delegating: in large groups where a full group deliberation is infeasible, people choose delegates who they trust to represent their views. Delegates can further delegate, so that you can defer the decision of who to delegate to.
For decisions which do not realistically require consent from everyone, EG choosing a restaurant when not everyone realistically will go, a kickstarter-like mechanism for tabulating how many would go along with which decision. Some kind of utilitarian voting technique to go along with this, helping to select the best out of the options which have enough potential support.
Approval voting, but everyone states the percentage approval at which they would defer to the group decision. This allows us to shift between requiring full consensus vs simply the highest total approval number (even if that’s a tiny minority).
A “Bayesian Database” of all scientific information, in which hypotheses are registered with descriptions in a formal language (so that we can apply a description length prior), and submitted data automatically updates all the hypotheses.
The previous but also connected to a prediction market somehow.
The previous but also with argument mapping capability, including appropriately propagating information across probabilistic arguments.
More generally, prediction markets with argument mapping.
The previous, plus rewarding valid arguments thru picking up any arbitrage implied.
A prediction market which also has a connected way to put up money to make events happen. The prediction market is used to solve the credit-assignment problem, and also solve it predictively, so the market selects the most efficient allocation of your money it can find, including paying some up-front and also paying bonuses later based on actual perceived effectiveness.
A quadtratic funding effective altruist charity fund, with a connected prediction market to rank charities (just for informing givers, not with any strict connection to money allocation), of course including feedback about what actually happens with grants. The prediction market uses its own virtual currency (so it’s not easily manipulated by outside interests, and just finds the best forecasters). [A problem with this is that prediction markets aren’t going to be particularly good for evaluating x-risk causes.]
An effective altruist / rationalist version of LinkedIn/Klout/etc: EAs/rationalists rate each other for various important properties, using overlapping webs of trust, with (hopefully) Bayesian probabilistic soundness in how trust inferences are propagated. This helps recruitment and hiring for EA orgs, and facilitates finding partners for less formal collaborations, etc.
A microloan/microgrant fund for EAs, with some sort of accountability, bringing us closer to just having one big pot of money which all EAs can draw from as needed. (A “loan/grant” must be “repaid”, but not necessarily with money; IE it can be “repaid” with some sort of impact certification?)
A phone app that alerts you when you are in close physical proximity to another rationalist, eg in an unfamiliar city.
A heat map of density of rationalists, including eg which restaurants they frequent.
An app for coordinating rationalist/ea group houses, allowing you to put in housing preferences (people you’d like to live with, people you wouldn’t live with, space requirements, rent cap, requirements for space, other requirements for the house, distance from work, etc) and jointly optimizing everything to find good proposals.
Like Github, but for EA projects: a single place with a bunch of projects, descriptions of how to contribute, tools for managing tasks (similar to bug report tickets), etc.
A way to classify claims made by pundits/journalists/writers/etc in articles or public statements, and then record follow-up fact checking / prediction accuracy, to essentially force accountability on them, and help people estimate the accuracy of new statements from the same sources.
A way to implement efficient Futarchy for small groups, on par with how easy it is to run small groups via democratic vote.
Like voting, but with more explicit rules for good deliberation and delegation of deliberation.
Rather than voting, deliberate until unanimous consensus is reached, forcing the group to address minority concerns.
Add to the previous the idea of delegating: in large groups where a full group deliberation is infeasible, people choose delegates who they trust to represent their views. Delegates can further delegate, so that you can defer the decision of who to delegate to.
For decisions which do not realistically require consent from everyone, EG choosing a restaurant when not everyone realistically will go, a kickstarter-like mechanism for tabulating how many would go along with which decision. Some kind of utilitarian voting technique to go along with this, helping to select the best out of the options which have enough potential support.
Approval voting, but everyone states the percentage approval at which they would defer to the group decision. This allows us to shift between requiring full consensus vs simply the highest total approval number (even if that’s a tiny minority).
A “Bayesian Database” of all scientific information, in which hypotheses are registered with descriptions in a formal language (so that we can apply a description length prior), and submitted data automatically updates all the hypotheses.
The previous but also connected to a prediction market somehow.
The previous but also with argument mapping capability, including appropriately propagating information across probabilistic arguments.
More generally, prediction markets with argument mapping.
The previous, plus rewarding valid arguments thru picking up any arbitrage implied.
A prediction market which also has a connected way to put up money to make events happen. The prediction market is used to solve the credit-assignment problem, and also solve it predictively, so the market selects the most efficient allocation of your money it can find, including paying some up-front and also paying bonuses later based on actual perceived effectiveness.
A quadtratic funding effective altruist charity fund, with a connected prediction market to rank charities (just for informing givers, not with any strict connection to money allocation), of course including feedback about what actually happens with grants. The prediction market uses its own virtual currency (so it’s not easily manipulated by outside interests, and just finds the best forecasters). [A problem with this is that prediction markets aren’t going to be particularly good for evaluating x-risk causes.]
An effective altruist / rationalist version of LinkedIn/Klout/etc: EAs/rationalists rate each other for various important properties, using overlapping webs of trust, with (hopefully) Bayesian probabilistic soundness in how trust inferences are propagated. This helps recruitment and hiring for EA orgs, and facilitates finding partners for less formal collaborations, etc.
A microloan/microgrant fund for EAs, with some sort of accountability, bringing us closer to just having one big pot of money which all EAs can draw from as needed. (A “loan/grant” must be “repaid”, but not necessarily with money; IE it can be “repaid” with some sort of impact certification?)
A phone app that alerts you when you are in close physical proximity to another rationalist, eg in an unfamiliar city.
A heat map of density of rationalists, including eg which restaurants they frequent.
An app for coordinating rationalist/ea group houses, allowing you to put in housing preferences (people you’d like to live with, people you wouldn’t live with, space requirements, rent cap, requirements for space, other requirements for the house, distance from work, etc) and jointly optimizing everything to find good proposals.
Like Github, but for EA projects: a single place with a bunch of projects, descriptions of how to contribute, tools for managing tasks (similar to bug report tickets), etc.
A way to classify claims made by pundits/journalists/writers/etc in articles or public statements, and then record follow-up fact checking / prediction accuracy, to essentially force accountability on them, and help people estimate the accuracy of new statements from the same sources.