Please submit ideas for interesting markets and I will pay you to create, no strings attached (other than the markets must actually be created by the party being paid).
For the types of market topics that I have created, which might or might not be ones that appeal to me to fund, check out here:
https://manifold.markets/PatrickDelaney?tab=questions
Of course any market topic might be appealing to me to fund.
That being said, I am going to be much more likely to fund markets that are:
Verifiable by some third-party source and as unambiguous as possible. For example, if the topic is AI-related, what is the numeric, empirical benchmark being used, who publishes it, where can we follow it, and what is the threshold in order for the market to resolve? Laying all of these out unambigiously ahead of time in your idea description will help.
New, interesting ideas, not trying to be intellectually edgy, just trying to look into an interesting area, and if you're not confident that the market would be a good one to create because it might not be popular, go ahead and submit it and I'll fund it and then you don't have to worry about whether you're going to get an ROI.
Ideas that appeal to you personally and that you're likely to pay attention to on a regular basis and run well.
Lower net worth Manna individuals encouraged to apply. If you're already doing well in terms of Manna, I'm less likely to fund you. However you could post your idea her and say that you want to gift it to someone, and then a lower net worth individual could choose to create it and I will award them.
YES/NO, Multiple Choice, markets will take precident over more expensive markets to run, but this does not preclude other types of markets from being funded by any means.
There is no obligation, no strings attached.
If anyone chooses to pay me Manna to help fund this effort, I can tack on more.
If anyone wants to pay me Manna as a thanks after the fact or down the road, that's great, thank you.
If anyone wants to link back to this market from their description to show how it was sponsored, great, that might help others create additional markets as well, but no obligation to do so.
Please specify if you do send Manna where it should go, if you want it reinvested into this market.
Please link your created markets within the comment threads after you have created them.
1,000
3.00Coming back for a double-dip.
"Will a TinyBox ship in 2023?"
George Hotz recently launched tiny corp. Their product will be the tinybox, a $15k machine capable of "[running] 65B FP16 LLaMA out of the box". Based on WayBack Machine snapshots, the timeline for shipping the tinybox is 2-6 months, starting May 25 2023.
This question probes the progress of tiny corp, the tinybox, and the tinygrad library.
"How many former colonies will England play in the 2023 FIFA Women's World Cup?"
In the knockout round, England has the opportunity to play, Nigeria, Jamaica, Australia, and the United States. (They played no former colonies in the group stage.)
The question will have four possible choices: 1, 2, 3, and 4. The question will require M100 in funding. This list on Wikipedia will be used to determine which countries count as "former colonies".
If funded, I intend to make a question that follows up on this angle about the ongoing and implied "what about other possible LK-99-likes?" questions, so that people can arbitrage it properly (and also draw the appropriate mental connections as to "wait, what is LK-99 exactly?" and thus further perfect the distributed reasoning about things). I may focus on the "copper or gold?" detail recently discussed by Lai et alia in https://arxiv.org/pdf/2307.16040.pdf
Whoops, meant to post this in Conflux's thread instead, was going to post something else here.
This is an awesome grants program, thanks Patrick for creating it! I'm running my own, with fairly similar criteria (I didn't actually realize this program existed when I started mine...): differences are
no guideline about a third-party source (I just require general clarity)
I also offer optionally to make the market myself crediting you (if you feel the market would receive more attention/legitimacy this way)
limited to users with net worth <M$2000 (but this is negotiable)
Anyway, if you think your market is a better fit for my grant program than Patrick's, I'll drop the link to mine:
Will a large language model be trained by Dec 31 2024 for the following task: parse history books or similar material to identify potential natural experiments. By natural experiments I mean configurations of the kind leveraged by econometrics to determine causal relations between variables, using techniques such as diff in diff, regression discontinuity, etc.
Will resolve based on the appearance of a convincing account of training either on arxiv, hal, or ideas.
The question: Will there be at least 12 "Dark Brandon" themed items on the Joe Biden campaign store by the end of 2023? Decided by how many items are in the "Dark" section of the store (https://shop.joebiden.com/dark/)
An idea that popped into my head when reading the discussions of UK media landscape. Couldn't find a market on Manifold so I made one. Thank you for your consideration!
https://manifold.markets/SarkanyVar/will-the-bbc-licence-fee-be-axed-by
