I think these are fine hypotheses about the tradeoffs here, though I disagree with most of the analysis. I have thought and read a lot about it, since like, my primary job is indeed to handle these exact tradeoffs and to build successful platforms here, but this current thread doesn’t seem like the right context to dig into them.
As one point, I think Manifold’s basic business model is “take a cut of the trading profits/volume/revenue”. The best alternative business model is “have people pay for finding out information via subsidies for markets”.
I don’t think Manifolds business model relies on advertisers or foundations. I think it scales pretty well with accuracy and usefulness of markets.
In the “take a rake of trading volume” model without any significant exogenous money coming it, there have to be enough losses to (1) fund Manifold, and (2) make the platform sufficiently positive in EV to attract good forecasters and motivate them to deploy time and resources. Otherwise, either the business model won’t work, or the claimed social good is seriously compromised. In other words, there need to be enough people who are fairly bad at forecasting, yet pump enough money into the ecosystem for their losses to fund (1) and (2). Loosely: whales.
If that’s right, the business rises or falls predominately by the amount of unskilled-forecaster money pumped into the system. Good forecasters shouldn’t be the limiting factor in the profit reaction; if the unskilled users are subsidizing the ecosystem enough; the skilled users should come. The model should actually work without good forecasters at all; it’s just that the aroma of positive EV will attract them.
This would make whales the primary customers, and would motivate Manifold to design the system to attract as much unskilled-forecaster money as possible, which doesn’t seem to jive well with its prosocial objectives. Cf. the conflict in “free-to-play” video game design between design that extracts maximum funds from whales and creating a quality game and experience generally.
I disagree with this. I think the obvious source of money for a prediction platform like Manifold is from people who want to get accurate information about a question, who then fund subsidies which Manifold gets a cut off. That’s ultimately where the value proposition of the platform comes from, and so where it makes sense to extract the money.
I read your comment as “have people pay for finding out information via subsidies for markets” being your “alternative” model, rather than being the “take a cut of the trading profits/volume/revenue” model. Anyway, I mentioned earlier why I don’t think being “controversial” (~ too toxic for the reputational needs of many businesses with serious money and information needs) fits in well with that business model. Few would want to be named in this sentence in the Guardian in 2028: “The always-controversial Manifest conference was put on by Manifold, a prediction market with a similarly loose moderation norms whose major customers include . . . .”
I think it’s a minor issue that is unlikely to drive anyway who actually has a “hair-on-fire” problem of the type that a prediction market might solve. I am confident anyone with experience building internet platforms like this would consider this a very irrelevant thing to worry about at the business stage where Manifold is at.
I think these are fine hypotheses about the tradeoffs here, though I disagree with most of the analysis. I have thought and read a lot about it, since like, my primary job is indeed to handle these exact tradeoffs and to build successful platforms here, but this current thread doesn’t seem like the right context to dig into them.
As one point, I think Manifold’s basic business model is “take a cut of the trading profits/volume/revenue”. The best alternative business model is “have people pay for finding out information via subsidies for markets”.
I don’t think Manifolds business model relies on advertisers or foundations. I think it scales pretty well with accuracy and usefulness of markets.
In the “take a rake of trading volume” model without any significant exogenous money coming it, there have to be enough losses to (1) fund Manifold, and (2) make the platform sufficiently positive in EV to attract good forecasters and motivate them to deploy time and resources. Otherwise, either the business model won’t work, or the claimed social good is seriously compromised. In other words, there need to be enough people who are fairly bad at forecasting, yet pump enough money into the ecosystem for their losses to fund (1) and (2). Loosely: whales.
If that’s right, the business rises or falls predominately by the amount of unskilled-forecaster money pumped into the system. Good forecasters shouldn’t be the limiting factor in the profit reaction; if the unskilled users are subsidizing the ecosystem enough; the skilled users should come. The model should actually work without good forecasters at all; it’s just that the aroma of positive EV will attract them.
This would make whales the primary customers, and would motivate Manifold to design the system to attract as much unskilled-forecaster money as possible, which doesn’t seem to jive well with its prosocial objectives. Cf. the conflict in “free-to-play” video game design between design that extracts maximum funds from whales and creating a quality game and experience generally.
I disagree with this. I think the obvious source of money for a prediction platform like Manifold is from people who want to get accurate information about a question, who then fund subsidies which Manifold gets a cut off. That’s ultimately where the value proposition of the platform comes from, and so where it makes sense to extract the money.
I read your comment as “have people pay for finding out information via subsidies for markets” being your “alternative” model, rather than being the “take a cut of the trading profits/volume/revenue” model. Anyway, I mentioned earlier why I don’t think being “controversial” (~ too toxic for the reputational needs of many businesses with serious money and information needs) fits in well with that business model. Few would want to be named in this sentence in the Guardian in 2028: “The always-controversial Manifest conference was put on by Manifold, a prediction market with a similarly loose moderation norms whose major customers include . . . .”
I think it’s a minor issue that is unlikely to drive anyway who actually has a “hair-on-fire” problem of the type that a prediction market might solve. I am confident anyone with experience building internet platforms like this would consider this a very irrelevant thing to worry about at the business stage where Manifold is at.