To make a long story very short: my friends and I started a company, it failed, but we have a decent amount of money left over (~100kish). We talked it over with our investors, and both of us want to use the remaining funds to chase a new idea.
Right now, our team is in the discovery phase of picking a completely new direction to go in. We're researching tons of different routes, and because of my already-established interest in prediction markets, I was tasked with scoping out the space for any leads on a potential idea. (Apparently, prediction markets are "hot", according to our investors...who knew my fake-internet-money gambling addiction might pay professional dividends?)
So far, I've found lots of interesting ideas on potential use cases of prediction markets, from things like policy markets or replication markets to solving two-sided adverse selection across a ton of different areas like grants programs, employee hiring, and venture funding.
I love these ideas, but startup 101 (and partially the reason we failed on our first shot) dictates that you can't build towards an idea, you must build to solve a specific problem. Additionally, it's really hard with our limited funding to build out consumer-facing applications that require a critical mass of users to turn profitable. So, I ask everyone who reads this:
What problems exist in the prediction markets space where the form factor of a solution could be a company?
Where are the gaps in wider market infrastructure that need to be filled?
Who, if anyone, needs anything built that a handful of twenty-somethings with $100k could build?
I'll reward bounties for anything I deem helpful. The questions are specifically about prediction/betting markets, but feel free to drop an answer if you have a problem of any type that we might be able to solve. If there's anything you don't want to talk about publicly, feel free to reach out via email at jasuch@uw.edu.
1,000
1.00Just an idea I've been thinking about - if one wanted to make prediction markets a public fixture, I'd go for a PMaaS approach. Not necessarily PMs for internal forecasting within businesses, but rather cooperating with media. I guess you've seen those polls on websites of media outlets ("Biden or Trump"?) - what if you could partner with media outlets to replace them with prediction markets? Now, I have no idea whether it would work and what benefits exactly it would give to the outlet, but there's research suggesting e.g. quizzes both help users learn and benefit the outlet. Similar tests with prediction markets seem very interesting, if you can make the user experience simple enough...
Not prediction markets (albeit maybe in the category of truth-seeking), but I'm building https://n1.tools and currently considering next steps. Even if it doesn't turn into a collaboration, I'd be down to chat about your own startup experience!
Random idea that I just came up with but which I would unironically support: a publication about the future (vaguely like the Base Rate Times, but more spruced up). It would have articles about various topics that people could look up (eg Massachusetts, or AI, or Africa, or Jose Luis Ricon, or whatever), and those articles would discuss the future prospects of said topics in lurid detail...
but instead of citing the mainstream media like Wikipedia does, this website would cite individual prediction markets/questions (eg individual markets/questions on Manifold or Polymarket or Metaculus or such).
EDIT: oh, I almost forgot - people should be able to edit this in real time! (kind of like the anyone-can-bet paradigm of Manifold, or the anyone-can-edit paradigm of Wikipedia)
I’m relatively new to prediction markets, but a major source of pain seems to clarity in decision criteria. “Timely and accurate resolution guaranteed by XYZ” seems like a service that could exist. Ratings agencies exist in lots of sectors, but prediction markets may not be mature enough to support it.
How about a platform like Wikipedia but also featuring news, where users can rate and comment on whether an article is real or not? For instance, if someone posts an article about the Earth being flat, users can vote on whether it's true or not. There could be a ranking of the most controversial topics, or those where over 95% of users agree on a certain subject. The main idea would be a fact-checking platform where users, through votes, ratings, or something similar, verify whether a topic is real or false. It could also include articles with future predictions in any field. It's just a spur-of-the-moment idea. If more ideas come to me, I'll come back. Good luck!
The questions are specifically about prediction/betting markets
The part that appeals to me the most about prediction markets is calling people out on their bullshit, making people put their money where their mouth is, making them feel humiliated for believing in nonsense and hopefully increasing critical thinking skills and decreasing the amount of disinformation spreading around (and destroying our society).
Unfortunately, betting with real money is illegal in many places, and certain false claims aren't easy to turn into objective bets. But anything you can do in that space to make people put up or shut up and face real consequences for spreading disinformation/conspiracy theories/etc. would I think make society a better place. (And the wager could be something other than money that still has real consequences to stay legal.) Something that could be integrated into or tacked onto websites where people tend to post stupid shit like Twitter / Reddit / etc. would be good. Not sure the best way to do that.
Truthcoin was supposed to do this, for instance, but is basically defunct as far as I can tell.
but feel free to drop an answer if you have a problem of any type that we might be able to solve
I want more crowdsourced reviews of whether things work or not. Specifically, I really miss the website CureTogether, which gathered from users what symptoms they have for a particular condition, what treatments they've tried, and whether those treatments helped or made the problem worse. 23andme bought it and mothballed it, which is annoying, and I haven't found anything similar. Literally just cloning the functionality of that site so it's available again under a new name would be awesome.
But also other things in the same vein, like "what self-help books have you tried for your mental health problems and did they help or not". Book reviews kind of cover this, but … meh. I want more of a summary, a chart. Also, things that work for some people might not work for others, so collaborative filtering would be even better. ("Users like you found this treatment / book / fashion advice to be the most helpful, you might, too.")
Also trust networks to verify that a username on the internet is a real person, or reputation / credibility networks to verify that someone has skill in a particular area without having to toot their own academic credentials, stuff like that. It's been attempted a bunch of times before but nothing has taken off that I'm aware of.
(I have insomnia lol.)
Make a site where you can simulate buying and trading any manner of securities or assets. Allow users to benchmark their rate of return and see how they perform against the market. Allow people to crowdsource predictions on securities.
Basically Manifold but for the stock market? Place for risk-averse people to test out their investment strategies. I’ve been looking for a site like this and haven’t found a satisfactory one.