The biggest enemy of prediction markets is the market itself. That is, unless you have inside information or are in the far right tail of market participants, you have a negative expected value betting in an efficient market.
So then it makes sense that sports and entertainment are where the preponderance of dumb money goes. Many people are happy to lose money betting for entertainment value. Compare that to betting on the date OpenAI’s next AI model gets released, with OpenAI employees as your counterparty.
Political markets are the unique crossover where people want to bet for entertainment value (attracts enough dumb money to also attract smart money), and there is a valuable signal to observe (accurate sports betting lines aren’t *valuable* signals), and the value isn’t merely the leaking of inside information (is culture drifting towards “insider trading is good actually”? I think not).
Insider trading is the present norm wrt politics. If anything, prediction markets add competition there. Probably the largest reason government would try to regulate them too.
It would be interesting to see companies (or other entities) fund markets for things they value, as a way to make them positive-sum and draw in more smart-money participants. Like Google Opinion Rewards but for prediction markets.
Totally agreed that this is just the beginning of things to come as long as prediction markets avoid shooting themselves in the foot with exposure to controversial markets.
For what it's worth, I think they make much more sense and are less scummy than betting on sports through sportsbooks since the business model profits off volume, not setting profitable odds.
Wrote a pretty long deep dive on Kalshi's history, how it works, and what it could enable this week if anyone wants to check it out!
One of the best ways of popularising novel ideas is through fiction. Are there stories of worlds where prediction markets for governance are adopted (and ideally an important feature of the world) or have you considered writing one?
If there aren't and you have not, I think that you should do so or encourage others to do so.
Can we assume that the concomitant rise of predication markets and AI/LLM analysis will result in prediction markets primarily being dominated by the model developers themselves or just the side of the market working to write the best poetry to extract the more accurate predictions using the same models?
As a source of information to advise decision-makers, prediction markets have a new competitor: AI. It's plausible that in the next decade or two, machine learning models will become better than prediction markets at predicting elections and sports games and so on. The ML models will also be cheaper, a few seconds of GPU time vs thousands of humans betting. Though it will then be a while after that before the ML models are trusted as much as the humans.
If AI is the best way to answer a question, then you can make money by betting on the AI's answer to a prediction market. And thus the price will correct. Still canonical
The biggest enemy of prediction markets is the market itself. That is, unless you have inside information or are in the far right tail of market participants, you have a negative expected value betting in an efficient market.
So then it makes sense that sports and entertainment are where the preponderance of dumb money goes. Many people are happy to lose money betting for entertainment value. Compare that to betting on the date OpenAI’s next AI model gets released, with OpenAI employees as your counterparty.
Political markets are the unique crossover where people want to bet for entertainment value (attracts enough dumb money to also attract smart money), and there is a valuable signal to observe (accurate sports betting lines aren’t *valuable* signals), and the value isn’t merely the leaking of inside information (is culture drifting towards “insider trading is good actually”? I think not).
Insider trading is the present norm wrt politics. If anything, prediction markets add competition there. Probably the largest reason government would try to regulate them too.
It would be interesting to see companies (or other entities) fund markets for things they value, as a way to make them positive-sum and draw in more smart-money participants. Like Google Opinion Rewards but for prediction markets.
Totally agreed that this is just the beginning of things to come as long as prediction markets avoid shooting themselves in the foot with exposure to controversial markets.
For what it's worth, I think they make much more sense and are less scummy than betting on sports through sportsbooks since the business model profits off volume, not setting profitable odds.
Wrote a pretty long deep dive on Kalshi's history, how it works, and what it could enable this week if anyone wants to check it out!
https://open.substack.com/pub/theuncredentialed/p/kalshi-building-a-truth-machine?r=26ya72&utm_campaign=post&utm_medium=web
One of the best ways of popularising novel ideas is through fiction. Are there stories of worlds where prediction markets for governance are adopted (and ideally an important feature of the world) or have you considered writing one?
If there aren't and you have not, I think that you should do so or encourage others to do so.
Can we assume that the concomitant rise of predication markets and AI/LLM analysis will result in prediction markets primarily being dominated by the model developers themselves or just the side of the market working to write the best poetry to extract the more accurate predictions using the same models?
As a source of information to advise decision-makers, prediction markets have a new competitor: AI. It's plausible that in the next decade or two, machine learning models will become better than prediction markets at predicting elections and sports games and so on. The ML models will also be cheaper, a few seconds of GPU time vs thousands of humans betting. Though it will then be a while after that before the ML models are trusted as much as the humans.
Different AIs give different answers, just as with humans. These markets are forums for settling those differences.
If AI is the best way to answer a question, then you can make money by betting on the AI's answer to a prediction market. And thus the price will correct. Still canonical
Sure, but if AI is the best way to answer a question you might do away with the expense of the market entirely.