I fully agree with Robin Hanson: We have to put decisions front and center. Predictions and prediction markets are just the means to make better decisions.
Hiring is not the killer application in my opinion. Think about a colleague Kim who got hired two years ago. can you imagine the company does a poll "was it a good decision to hire Kim two years ago?" Yes or No? They will then publish the result company-wide and resolve some prediction markets based on that. As a European, that surely violates one or more laws here.
Deadlines look much more likely to me. Looking at the business around Agile, there already is a market for forecasting tools. For example, the Scaled Agile Framework suggests a "confidence vote" at the big planning meeting every two months. Turning that into a prediction market is only slightly more technical. Atlassian could build it into Jira. It feels like we are so close.
We think fake news is an interesting use-case, although not convinced it necessarily is a great fit. Need to think more about it. Enjoyed your article, but it's worth mentioning that you picked an example where it worked, but that does not prove the whole. Moreover, you mention LK-99 as somehow being a positive, but to our recollection, PMs stayed far more bullish on its feasibility for far longer than traditional media. Then again, we have talked with Saar Wilf from Rootclaim on our podcast in the past, so we do think there is value in applying probabilistic thinking to uncertain and controversial events: https://www.thirdimage.media/saar-wilf-rootclaim-interview/
When it comes to hiring markets, I think the one you described certainly would run into problems. But, can they not be constructed in a way that is far less controversial? Instead of "was it a good decision to hire Kim", is it "will Kim achieve X, Y, Z KPIs" or use a prediction poll instead of market and keeping aggregate numbers away from everyone but HR / relevant hiring managers. Also, your market is ex post hiring whereas the ones Hanson is talking about are ex ante markets.
All that being said, we agree that deadlines has the clearest and least controversial path towards adoption. Although that will differ depending on what industries / types & sizes of companies we are dealing with.
Unfortunately, I doubt that this newsletter will make you even minimum wage, which is unfortunate because you two are so good at this. Problems: too narrow a news focus, and the overall falling trend of news media of all sorts.
The only contrary data I have found so far is The Bulwark https://www.thebulwark.com/, growing vigorously while still, like you, using Substack as its base platform. It includes in-person events, videos including live video events, and several newsletters. You two both have considerable talents for video and come across well in person, so this might be a way for you to buck the trends.
That said, I suspect The Bulwark still is losing money while growing. According to what I have read online, it is bolstered by donations even though it is not a nonprofit.
Thanks to Open Philanthropy, huge amounts of money have been donated to crowd-sourced forecasting entities. Now the spigot has turned off, and its allied EA is hemorrhaging participants and donations. Much if not all of this is in danger of clawbacks in the FTX bankruptcy case. I hope you are not at risk! Given the world of increasing hurt on your field, you might gain traction.
We are still working on BestWorld, currently conducting research into ways to automate creating a news aggregation system with a future emphasis, using crowdsourced forecaster rationales + stated probabilities as inputs. Clearly both journalism and crowdsourced forecasting are money-losing propositions, with rare exceptions on the journalism side only. So perhaps we are chasing a mirage.
Aw, thanks Carolyn. Glad to hear things are still progressing on BestWorld. We ought to catch up one of these days -- maybe even go for horse riding part 2! 😂
But yeah, your final thought that "perhaps we are chasing a mirage" is one that's been in our heads for a while and Hanson's talk has only furthered it. Would love to hear a bit more about how you are thinking about this idea, as well as Hanson's address.
We are thinking about more interesting things to do and find ways to gain larger traction, some of which you'll hopefully see quite soon!
It's good to hear about your quest for seeking the better good for humanity.
Our mutual friend, Miles Palmer, has managed to make tons of money while only funding technologies that help humans and the rest of our biosphere. He is an example of how to do well by doing good: a combination of brains and ethics and knowing how to work together (the hard part!)
At Manifold, we took Hanson's message to heart. The high value use case we decided to try out first is forecasting long term relationships — in part because we know from experience that people really love trading on markets about relationships. No this is not a joke haha!
In the last two weeks, we've built a whole dating app to do this. I'm about to announce the launch of our MVP later today.
Give it a shot: https://manifold.love (You can also participate as a matchmaker and not a dater.)
I actually have other ideas that I think could be good, including a consumer spin on the forecasting which candidates will succeed at companies.
Have you subscribed to the newsletter? 😉 The best I can give you is a clue. The tool we use is also the name of a sorcery from Guilds of Ravnica with CMC >4 and one of its colors is green. Let me know if you need more hints!
I'm stumped on the name. There are only 7 options, and none of them seem super on flavor for such a tool. ("Find" and "Flourish" are the closest matches.) I tried googling them anyway, but nothing turned up.
I think the impulse to apply forecasting to a corporate application may be too limiting. I like the ideas around fake news checking.
But it occurs to me that someone promoting “fake news” might really believe that what they are pushing is the truth. And the other guy is the one who’s really pushing fake news.
I think that’s a pretty good summation of the state of our society.
Maybe a better approach than discrediting “fake news” (which is an inherently biased statement because it presupposes that you already know which side is telling the truth) would be to set up a system of forecasting questions designed to test which narrative comes closer to predicting measurable outcomes.
- Did the Population Bomb result in millions of deaths by famine by 1985?
- Did Defund the Police improve criminal outcomes in places where it was tried?
- Will Argentina’s new President reduce inflation there?
You get the point.
The way I have been seeing forecasts used seems more likely designed to further a particular narrative than to judge between competing narratives.
I fully agree with Robin Hanson: We have to put decisions front and center. Predictions and prediction markets are just the means to make better decisions.
Hiring is not the killer application in my opinion. Think about a colleague Kim who got hired two years ago. can you imagine the company does a poll "was it a good decision to hire Kim two years ago?" Yes or No? They will then publish the result company-wide and resolve some prediction markets based on that. As a European, that surely violates one or more laws here.
Deadlines look much more likely to me. Looking at the business around Agile, there already is a market for forecasting tools. For example, the Scaled Agile Framework suggests a "confidence vote" at the big planning meeting every two months. Turning that into a prediction market is only slightly more technical. Atlassian could build it into Jira. It feels like we are so close.
To contribute another idea: Fake news and fact checking seems to work well if the platform allows to setup prediction markets quickly. https://marketwise.substack.com/p/the-battle-against-fake-news-how
We think fake news is an interesting use-case, although not convinced it necessarily is a great fit. Need to think more about it. Enjoyed your article, but it's worth mentioning that you picked an example where it worked, but that does not prove the whole. Moreover, you mention LK-99 as somehow being a positive, but to our recollection, PMs stayed far more bullish on its feasibility for far longer than traditional media. Then again, we have talked with Saar Wilf from Rootclaim on our podcast in the past, so we do think there is value in applying probabilistic thinking to uncertain and controversial events: https://www.thirdimage.media/saar-wilf-rootclaim-interview/
When it comes to hiring markets, I think the one you described certainly would run into problems. But, can they not be constructed in a way that is far less controversial? Instead of "was it a good decision to hire Kim", is it "will Kim achieve X, Y, Z KPIs" or use a prediction poll instead of market and keeping aggregate numbers away from everyone but HR / relevant hiring managers. Also, your market is ex post hiring whereas the ones Hanson is talking about are ex ante markets.
All that being said, we agree that deadlines has the clearest and least controversial path towards adoption. Although that will differ depending on what industries / types & sizes of companies we are dealing with.
Outstanding! Both well sourced and well written!
Unfortunately, I doubt that this newsletter will make you even minimum wage, which is unfortunate because you two are so good at this. Problems: too narrow a news focus, and the overall falling trend of news media of all sorts.
The only contrary data I have found so far is The Bulwark https://www.thebulwark.com/, growing vigorously while still, like you, using Substack as its base platform. It includes in-person events, videos including live video events, and several newsletters. You two both have considerable talents for video and come across well in person, so this might be a way for you to buck the trends.
That said, I suspect The Bulwark still is losing money while growing. According to what I have read online, it is bolstered by donations even though it is not a nonprofit.
Thanks to Open Philanthropy, huge amounts of money have been donated to crowd-sourced forecasting entities. Now the spigot has turned off, and its allied EA is hemorrhaging participants and donations. Much if not all of this is in danger of clawbacks in the FTX bankruptcy case. I hope you are not at risk! Given the world of increasing hurt on your field, you might gain traction.
More bad news: the entire news media ecosystem is failing fast. See https://unprecedented.ghost.io/archive/news-crash/ for a deep dive into the data. See also the imploding display ad ecosystem here: https://www.therebooting.com/p/newsletter-moneyball
We are still working on BestWorld, currently conducting research into ways to automate creating a news aggregation system with a future emphasis, using crowdsourced forecaster rationales + stated probabilities as inputs. Clearly both journalism and crowdsourced forecasting are money-losing propositions, with rare exceptions on the journalism side only. So perhaps we are chasing a mirage.
Aw, thanks Carolyn. Glad to hear things are still progressing on BestWorld. We ought to catch up one of these days -- maybe even go for horse riding part 2! 😂
But yeah, your final thought that "perhaps we are chasing a mirage" is one that's been in our heads for a while and Hanson's talk has only furthered it. Would love to hear a bit more about how you are thinking about this idea, as well as Hanson's address.
We are thinking about more interesting things to do and find ways to gain larger traction, some of which you'll hopefully see quite soon!
It's good to hear about your quest for seeking the better good for humanity.
Our mutual friend, Miles Palmer, has managed to make tons of money while only funding technologies that help humans and the rest of our biosphere. He is an example of how to do well by doing good: a combination of brains and ethics and knowing how to work together (the hard part!)
At Manifold, we took Hanson's message to heart. The high value use case we decided to try out first is forecasting long term relationships — in part because we know from experience that people really love trading on markets about relationships. No this is not a joke haha!
In the last two weeks, we've built a whole dating app to do this. I'm about to announce the launch of our MVP later today.
Give it a shot: https://manifold.love (You can also participate as a matchmaker and not a dater.)
I actually have other ideas that I think could be good, including a consumer spin on the forecasting which candidates will succeed at companies.
What program is that screenshot of a list of predictions on different platforms from?
Have you subscribed to the newsletter? 😉 The best I can give you is a clue. The tool we use is also the name of a sorcery from Guilds of Ravnica with CMC >4 and one of its colors is green. Let me know if you need more hints!
I am now!
I'm stumped on the name. There are only 7 options, and none of them seem super on flavor for such a tool. ("Find" and "Flourish" are the closest matches.) I tried googling them anyway, but nothing turned up.
Aw shucks, you did get it right! Flourish! https://flourish.studio
I think the impulse to apply forecasting to a corporate application may be too limiting. I like the ideas around fake news checking.
But it occurs to me that someone promoting “fake news” might really believe that what they are pushing is the truth. And the other guy is the one who’s really pushing fake news.
I think that’s a pretty good summation of the state of our society.
Maybe a better approach than discrediting “fake news” (which is an inherently biased statement because it presupposes that you already know which side is telling the truth) would be to set up a system of forecasting questions designed to test which narrative comes closer to predicting measurable outcomes.
- Did the Population Bomb result in millions of deaths by famine by 1985?
- Did Defund the Police improve criminal outcomes in places where it was tried?
- Will Argentina’s new President reduce inflation there?
You get the point.
The way I have been seeing forecasts used seems more likely designed to further a particular narrative than to judge between competing narratives.