Any ideas for a concrete question that can estimate answers to our real question (the real question is "According to Quroe's video's 'energy accounting' or my vague 'utility accounting', are prediction markets net good or bad?")

Maybe something like... If the money/time/effort spent creating prediction markets and providing the initial liquidity, is more than the total loss (somehow measure loss to the gamblers), then it's a net good? So like, we could proxy this by measuring the total mana invested as liquidity in all markets, vs the balance loss of all users whos balances went down... or something?

Those numbers don't actually answer anything, but maybe there's some other measurable set of numbers that does. Any ideas?

@DannyqnOht Is there really any good way to quantify it, though? Net financial gain to society? (e.g., money lost by addicted gamblers vs money saved via more accurate worldview from prediction markets)

But gambling is addictive for humans. And has a pretty tragic history.

We can't keep designing the entire society to accommodate the bottom 25% of humanity.

@skibidist hmmmm.... I think I'm just gonna flat disagree with you on that one. In terms of both fact and morals.

1) The bottom 25% includes things like 'lost an arm in a car crash that was someone else's fault'. Maybe like... bottom 1% or something could do what you're thinking? But 25% is wayyyy too much.
2) I think economic growth will still outpace the accumulation of broken genes for long enough that we can actually accomplish it, actually pretty easily, for the bottom 25%. We've only got like 2 or 3 more generations max before the singularity anyway. Plus, bottom 25% of the world produces way more than enough to cover their costs at the energy/utility level, so the implication that "we" are paying for "them" is just incorrect.
3) To borrow a morality test from yudkowsky: I think if you didn't know who you were gonna wake up as, and you had to pick a society to poof into, you wouldn't pick a society that did that. Not without some really convincing argument that culling the bottom 25% caused extreme gains for the ones who lived. Knowing that you'd only get 75% of those gains in expectation. This whole argument really seems to me like it gets into eugenics. Let's not get into that.

To reinterpret this as graciously as possible... Maybe you're saying we shouldn't accommodate people with gambling addictions?

Sorry, speaking as someone who's made great returns on prediction markets so far and uses a lot of math to do it, it just seems to easy to me to imagine another world where I was addicted. I just can't see gambling addiction in particular as something we should just flat ignore as a potential cost, any more than a competent advertiser should ignore sex appeal and say "that's just irrationality"; humans are wired like humans, and we should take that into account when deciding what to do.

I could see arguments that we don't have a right to restrict anyone's access. But when deciding what to champion as a good move for society, I think I need to take into account how it will affect the bottom 25% too. Not focus only on them, for sure! Gains of 100 utils for everyone in the top 10% can offset losses of 4 utils for everyone in the bottom 25%! But you have to do the math, or otherwise argue the gains exceed the losses; you can't just decide to ignore a whole section.

@skibidist why not go all the way and weed out all the Untermensch. The weak burden the strong, don't they?

@ItsMe For one, because it would violate the Golden Rule

I think if you didn't know who you were gonna wake up as, and you had to pick a society to poof into, you wouldn't pick a society that did that

@DannyqnOht I would, or it would not be me. I would rather keep reincarnating than live 100 lives as some low-life. I don't buy your utilitarian calculations because nothing below top excellence has any utility to me. Thank you for your attention to this matter.

@skibidist that was a quick 180. And it only took one German word.

@ItsMe from what to what?

@skibidist if you're in the top 75%, you shouldn't need me to spell it out.

@ItsMe i am bottom 25% when it comes to understanding liberal

@skibidist clearly you are the liberal, given your reluctance to preserve the Ubermensch.

"TODO fill in examples" lol

It's interesting gambling, much more so than a slot machine.

@AlanTennant This brings up a good point. I shouldn't be asking if they'll cause harm to people. I should be asking if they'll cause more harm than the alternative (existing slot machines).

I think... it probably will? I think prediction markets, and the act of championing prediction markets as valuable, will lead to more gambling addictions than there would be without them.

Also, aside, I think Manifold is so far ahead of the pack on the morality question. I think it's more of a question of, are the people running kalshi/polymarket/others (or the people making legal decisions around this in the US) taking it so far in the wrong direction, that if you had to choose all prediction markets or none, should we choose none?

@DannyqnOht Before the conversation breaks into a larger topic, I would like to offer this video essay to the conversation first. I would like to adopt the thesis it presents as something I believe in now too. https://youtu.be/c1zssAsg9ZY?si=fKzQDtkQ7emjDYdw

Un-spoiler this once you have seen the video:

I would like to offer these questions to you.

• What are you getting out of prediction markets?

• Do you think the collated information of the markets is being used in productive ways that benefit greater systems?

• Do you think the answer to the previous question depends on what any individual prediction market itself is about?

I want to continue this conversation into play money versus real money prediction markets too, but I think starting with these questions is more fundamental.

@Quroe I appreciate that - these are also the questions I wanted to think about, but you cut through a ton of ambiguity with that video!

Instead of energy as the thing to track and optimize, as in your video, I've had an intuition around using individually-defined-utility as a the objective; with some way to aggregate it across lots of individuals; and I'm kinda assuming that aggregation mechanism has some way to rescale everything so all agents (for now, could be all humans, depends if you allow 'groups of humans' to be an agent - churches? corporations? governments? In the future bot agents should be include too. Gets really complex if we include other species.) have a sorta equal voice when saying how much a given outcome affects total utility under your aggregation function. Ignore all the details for now. I was assuming log dollars was a good first-try proxy for individual utility, and while I think there's so many problems with this view, I think it's an easier & slightly more useful starting point than tracking 'energy under your command' (though energy-under-your-command is probably a more useful estimate than raw dollars).

So, to restate the question in terms of that video's language:
Prediction markets have some net-positive effects (like photosynthesis) and some net-negative effects (like seagulls stealing.) When we add them up, are they overall a net positive or a net negative? We don't often discuss the net negative effects. One obvious, but small, one is concentrating wealth in a way that probably net reduces total energy in the system. A system with 3 individuals who each have 10, 10, 10 probably has more aggregate utility than a system with 1, 1, 28 (this is a silly point. Lots of things aggregate wealth, even an optimal energy-making system will do that). A less silly net energy cost is, as currently implemented in Kalshi/Polymarket, and as sorta-hinted-at in the video that sparked my crisis: providing a breeding ground for extremely large scale corrupt behavior.

To answer your questions:
- I don't care what I'm getting out of prediction markets for the purpose of answering this question; I want to look at the society level. I do care about it because it makes me SUPER biased. I'm getting money and having fun, so I'm very biased to say "nah it's fine, right?"
- This. This is the crux of my crisis of faith. Is the collective behavior being used net productively? Or, can we expect it to be soon if it isn't yet?
- Yes, I think it absolutely depends on what markets are being made. Prediction markets on sports or random number generators don't help society. Prediction markets on crop yields probably do. I've seen a few major effects directing the flow of what kinds of markets get created:
- Kalshi/polymarket have their own mini monopolies on what real-money markets get created. Because 1) it's complicated to implement a prediction market 2) regulations 3) first movers
- In Manifold, the major force influencing public market creation is "will people look at this market". You get more mana the more people want to bet, you get better odds, it all becomes about getting eyeballs on your market (so far so good, but we know how that force tends to play out in the long run, what with current society and ads and social media).
- Manifold has some clearly useful markets - individuals spending to get data they care about (eg bayesian will I be sick), public info on things that might help people or companies make everyday decisions (eg will there be a crash), ...

Subtopic: I'm really not sure about political markets.
- In the question monopolies (poly, kalshi) it seems clear that only allowing discussion of certain things (eg politics, or crafting market descriptions & resolution criteria so that odds don't match first-pass meanings) can be used for net loss to society. It's hard for me to quantify this effect, I have huge priors but I haven't thought of any way to estimate or measure impact on net utility (net energy)
-In the question of manifold, I'm worried about similar concerns arising. It's not so clear a link, but there could be influence from outside that bleeds into market creators. I would love to have... some kind of reasoning that indicates, at least, that there's some kind of incentive/pressure on manifold users that would at least somewhat oppose a hypothetical villain mastermind who controlled the rest of social media. If everyone's already 10% brainwashed, is that enough to make it so only the biased markets get eyeballs, and manifold stays in the trap?

Subtopic: Gambling
- Even if markets produce clearly useful public information, it's not clear that the cost of that information is less than the value. What if, hypothetically, in the course of making that information, we cause 10% of the population to descend into gambling madness, and they're lost to scammers forever? And the remaining 90% of us get an extra 10 bucks of utility because of it, on average? That's an extreme example to show that there should probably be some cost at which we say "no" to prediction markets. But where are we in reality? Are there any clear arguments that the losses to the people who aren't good at reasoning or who are easily addicted, are worth it, in exchange for the gains for the rest of the world? How could we even go about quantifying that? Are we allowed to say "no" to prediction markets, even if the people who will be hurt by them want them?

@Quroe Just FYI, if you reply I probably won't be back on again until tomorrow (I'm more free during the day and less during the evenings)

@DannyqnOht how do you have time to write so much

@skibidist I just have the time to write *shrug

@Quroe Thoughts?

It seems like you really believe the legacy left by prediction markets will be one of clarity and net benefit for everyone, as opposed to a legacy of regret and thinking of them as just another bubble. I would love to know what makes you think that.

@DannyqnOht I want to say, first off, that I think play money prediction markets are almost certainly a net good to society. Sure, there are corner cases to that, but generally speaking, I think treating Manifold as a social media platform with the core tenet of reality-driven truth above all is a net boon to society. Those who stick around here are exposed to new ideas that can stand critical scrutiny.

In a world where social media is driven by engagement and measured in how many "likes" somebody can get, mana is a good proxy of that -- mana are your "likes". You get more "likes" of mana if you are knowledgeable (or know how to ask questions people find interesting).

Now, real money prediction markets are where things get sticky. I read a good article recently that really made me think about how Kalshi operates.

I also don't like scenarios like the ones shown in your originally referenced video that turn prediction markets about predicting whether something will happen or not into a bounty to make it happen -- especially when it advertises itself as predictive to the average trader instead of an effective bounty (like if a CEO will say a certain word in an earnings call, and they say the words or not as a favor to a business associate who made a bet one way or the other). I really like exploring this idea where I invite people to knowingly take part in this trap with my Nash Pit series. I think it is educational to make people really feel that whatever direction they trade makes it less likely that they will win that trade, because they just incentivise somebody to act against them with such trades.

(I intend to run a 3rd episode at Manifest this year if I get the opportunity. Episode 2 is "The Lost Episode" that has never happened yet at the time of comment, due to the unspecified trigger condition not being met yet to kick it off.)

So, with real money sites making a commission based on volume or trade action, it seems like you are weighing the effect of giving such platforms more power to do what they have been doing. You're also weighing if their actions as a whole have been more vicious or virtuous on balance -- is this correct? Or are you taking more of a Categorical Imperative approach to it?

Do you think minds are being changed by what they see on real money prediction markets? Are people learning lessons from being right/wrong on them? Or are we asking if gamblers are better off for visiting the casino?

@Quroe Yeah, you're right on the money with what I'm worried about. Do I give them more power to do what they're doing? Consequentialist for this discussion. I don't think gamblers are better off for visiting casinos.

This post is long. Summary:
- I wish prediction markets were this cool, good thing I have in my head.
- I thought we were building towards that, a bubble like the online poker one, and the end result would be the cool thing.
- Aside: fake money vs play money, uncertain how distinct they really are.
- Do I think we're getting there? I'm not sure...
- It would be sad to not take the money.
- Conclusion & next steps.

I... had this idea that the long-term value of prediction markets (hedging & insurance, future prediction, wisdom of the crowds) would eventually shine through against the backdrop of the gambling markets (which backdrop your video really emphasizes). The way Yudkowsky seems to think it'll eventually be in many of his books (I really liked how they were portrayed in this story he wrote, (short story, but a long unrelated read if you're just browsing) TL;DR: prediction markets are so ubiquitous and reliable that a nurse is shocked when one isn't available for a patient's prognosis. We should live in that world.) As I actually use Kalshi in reality though, I'm seeing more and more that it's just... people gambling. (More with real money than mana.) Even on things that matter, like 'will the government shutdown'. Gamblers betting on medical outcomes is not the win here; smart people being incentivized to provide information where it's needed most is the win here. First, I thought it was just an issue with the comment sections, and only the gamblers were posting comments, and the real bettors were silently profiting (because why give away your edge?). But over time... even the odds seem to be just straight irrational. Easily arbitrage-able. Not in huge ways though, so maybe it's just a liquidity issue... And maybe I'm just looking at the things that don't matter... but it has me really worried.

I was thinking it would be like the online poker bubble I remember, where during the bubble, tons of people could make money. Then the worst players would quit playing, and you had to have some skill. And then once the moderate-skill players were the bottom of the rung, only the really skilled people would make money and stick around. And finally, the bubble's done and it's only the best still playing. And I thought that, since prediction markets are supposed to add value to societies, it's worthwhile to participate in this bubble, especially if I can make a tool that makes everyone smarter, because there's value in that end-state. Unlike poker or sports betting.

But I'm really starting to doubt that mental model. What if it really is just straight gambling, and I'm just profiting off the gamblers, taking a cut from the casino? What if the larger effect of making a tool public, is to give more money to the corrupt house, or let the gamblers hang on two more days before they're broke; and the improvement in public information is negligible? Instead of getting closer to a world of accessible prediction markets, just ushering along to the public outcry that bans them.

(Aside: Excellent points about mana. It relates to my concern "both money and mana incentivize behavior/gambling". You made a very good point that mana is designed to incentivize what we want, in terms of a social-predictive-currency; as opposed to money which is much more of a currency designed to affect the real world. I think this extends into a good argument that we should, wherever possible, reduce the ability to use mana to change anything in reality. The major chink I'm not yet sold on (think I could be though), even though mana is well designed - does it have enough unintended side effects to be a problem, in the long run? I think manifold-the-website currently has an amazing culture in its users, but as a whole, we underestimate the long-term effect that mana-maximization-mindset will have on eroding that culture. I don't see people noticing often enough how their markets are incentivizing real world behaviors (your game show is sorely needed) - and new players who start taking advantage of this will get more "likes". It doesn't feel like there's been enough consideration of "where will this end up in the limit" kinda approach for me to fully commit. Without that kind of consideration, I hesitate to say "just making it play money will fix it." My assumption, before the limit consideration I'm hoping we do, is that eventually, if there's any way to convert mana to kiwis or donations to charity, mana will have the same end effect as money; and we should assume manifold will end up in the same place as real money markets.)

Back to the main point. Do I think minds are being changed? Are lessons being learned?
Maybe some.... I think (rather, hope) a lot of people will learn the lesson that lots of people have learned throughout history in casinos - don't bet for the fun of it, bet to win. But casinos are still around. If that argument will work to make prediction markets into good things, then why haven't casinos died out yet? Maybe humans brains are sufficiently similarly irrational, that there will always be lots of money to be made from odds-based betting; and because the architecture of the brain isn't going to change quickly, maybe that profit incentive will always dwarf the "give good information where it's needed" incentive.

With those answers... I don't know if just participating and profiting is the right thing to do, to be someone who helps give the world the benefits of good prediction markets. Maybe I should stop participating, at least on Kalshi (that's really sad. I was making so much more than my other investments... 😭) . That's my crisis of faith here.

It sounds like the stuff I'm talking about is getting pretty far from the topic you're interested in - it sounds like you're more interested in the play money, truth finding & visibility question; and I'm stuck on the real money, hurting the opposing party question. I think the bridge of "is making it play money sufficient to avoid the traps" might be worth discussing more until we get that hammered out, but I don't want to trap you into a discussion of the real money ethics if that's not something you care about. So. Maybe we should start with the bridge? Do some back and forth about whether the traps/problems I'm worried about are even things that need to be considered in Manifold? As opposed to diving straight into my concerns directly in a real-money environment. What do you think?

@DannyqnOht I am a sucker for philosophical topics, so rest assured: I am interested in talking about this. I am speaking about things that I have personal expertise/experience in, so I ask for forgiveness if my prism of experience distorts the topic you wanted to drive at.

The bubble you talk about is a novel idea to me. I think that financial tools typically are fad-like in their bubble stage, and I get the feeling we might be in that stage for prediction markets. Look at how NFTs fared, for example; look where the public zeitgeist is for those nowadays. I think society has a gambling problem right now, but real money prediction markets are more of a symptom than the root cause to me, as I see it.

Gambling has a way of sweeping up average people and pushing a technology over the critical mass to survive the barriers in the Law/Theory of Diffusion of Innovation (seen at the 11:08 mark of that link, but I enjoy the whole of that TED talk).

As for your tool, I think the people who will initially use it are the people who are already trading with an edge, and this will just make their edge more pronounced. If you are trying to spread wealth around equitibly, I think releasing your tool to the public will do the opposite of that, at least at first.

Maybe someday it gets adopted into an investment platform like Fidelity or Vanguard, and people will passively be able to use it without much thought or active guidance. It will just get packed into some index fund, and people will go on living their lives while the machine hums along.

Perhaps you're better off being the folklore representation of Robinhood (not the trading app) where you take your winnings from your tool and you give it to those in need through charity donation? It's a tough call.

Let's focus in on one of the questions you want to pose. Where do you want to take this topic, specifically? It can branch into a myriad of things easily, and there's a lot we could talk about, but let's narrow our scope on to one thing for now. Take the lead!

@Quroe Ok, yeah, let's talk about it concretely. Heck, maybe I'll convince you to be a playtester 😂

So, the idea was actually spurred from one particular sentence in that yudkowsky story above, about how there were experts giving probabilities to some program and then programmers & tools converting those probabilities into bets.

There's a lot more to prediction markets than "Market says X%, I think Y%, Y isn't X, so I bet."

Time value of money (eg the "Jesus returns" market that was nonzero), conditional information making the probability distribution have peaks in it, markets that are conditional on each other, Kelly criterion & managing risk, being a maker VS taker (thankfully not a problem on manifold 😂 ), managing fees...

So the point of the tool is to make it so that, you put in your odds, and it says "ok here's optimal betting if those odds are right". The major gains that are translatable to all prediction markets is that it can notice & take advantage of conditional dependence between markets (as long as you tell the tool about them); and it can actually bet optimally in a kelly sense (which studies show humans suck at, even in simple coinflip cases).

So.... That. It feels like the kinda thing that aught to make things better in the world overall? And the test version is making me good returns right now (though, if that's because of the math or because I have good input probabilities, who knows. The way I'm using it right now, is just as a confirm/deny when I want to bet on something, and it tells me appropriate scale.) But maybe all it does is make some gamblers overconfident, or makes them dribble their money out slower. Idk.

The tool is very proof-of-concept stage still, and very tied into Kalshi right now. One option would be to abstract a lot of it so it can work on Kalshi/Polymarket/Manifold, and then if I don't like one of them, I can just drop support for that market. But that sounds really hard to do.

But on the other hand... If we imagine it works, and it's Kalshi specific, and I release it... Am I just supporting evil?

(I don't have the money to actually release it, it's all running free tier right now. I'd have to add some kind of payment to pay for the servers if it actually got real usage, eg more than 100 users. But it's fun to imagine maybe it becomes a big thing, and figure out if I did something wrong in that case.)

Maybe I'm worrying about nothing here. You're right, it's gonna be superusers only pretty much; and the people using it aren't ones I need to worry about. And the idea isn't that original, I'm sure there's better stuff out there. So I probably have nothing to worry about.

But I kinda would like to feel a bit more confident when I emphatically tell my mom "no, I promise I'm not just encouraging degenerate gambling" 😂

@DannyqnOht I'll happily be a playtester. I have been a playtester for a few prominent things out there.

• I was a beta tester for Pokémon Go before it released to the public.

• You can see my efforts in how I'm trying to suggest improvements to Manifold.

• I have been asked to consult for some video game development projects by another Manifolder in the past.

• I have been asked to test user made tools developed to enhance the Manifold experience.

@Quroe I will probably take you up on that! Once I've like... swept, and taken out the trash, and stuff. Metaphorically.

But to the main point.

Is this a net bad?

Like, I have a kinda running hypothesis along the lines of:
- Humans, like most animals, are subject to classical conditioning
- The right stimuli in the right order trains us.
- Through dumb luck (or engineering) we can get a series of wins/losses that addicts us to gambling. Some people are more resistant or less, but we are all susceptible.
- With that, more eyeballs means more gambling addictions.

So I kinda am wondering... If I make a tool like that, am I net reducing gambling addictions? Or, what if, my tool accidentally smooths out the reward events, so even more people get addicted or something? Or, if I encourage everyone I know to make a betting account, and some of them get to hedge things they care about like hedging housing prices when they buy a house; but one of them gets addicted to gambling; have I done a net good or a net bad? What are the expected odds of those two outcomes? How would I even go about measuring them?

Thoughts on that (since you said you're willing to follow me down my preferred philosophical rabbit hole?)

@DannyqnOht Take a quick look at this article that @realDonaldTrump just wrote up. I think it is pertinent to this conversation. Thoughts?

@Quroe @realDonaldTrump Wow that seems like a pretty direct reply to what I'm talking about here 😂

I like a lot of the points! And want to compute the consequences of the theory! (will do so here)

(Aside: I may agree with the final points about it not mattering from a regulatory perspective, this won't affect the actual outcome; but think that for the purposes of deciding what policies you champion to the people around you everyday, our helplessness should not be a factor. We should think about what we'd do if we could control policy, when deciding what policies to champion - feels like part of the math required to figure out what policies are good or bad.)

Ok, so the theory. Let's start with my understanding of the post.

My summary: It's not about gambling or not, everything is gambling - valid point. Great concrete list of gambles was super helpful - (household purchases, like "do I buy solar panels", are gambles. Even "Will I use this blender" is a gamble.) Skipping over "gambles not backed by anything" just like the article did. Proposed in the article: We can rate the "quality" of a gamble from a social perspective. The post goes on to say "since prediction markets are better expected value & don't pay a house for the privilege of making your gambles backed by something, they're reasonably positive from a social perspective."

Okay. So my thoughts, and where I want to extend the theory.

0) I like the implied definition of gamble. Very clarifying.
1) I like the idea of rating gamble quality. I especially like the idea because it seems very plausible to demonstrate that giving agents in a game access to prediction markets, could strictly improve their combined expected value, relative to when they had to just make their own individual gambles. In concrete terms, if everyone had to decide whether to buy solar panels individually, I could see prediction markets being a provable improvement to everyone's payoff in that game. Still need to do the math, but it feels right.
2) I dislike the idea of using average expected utility as the only measure of gamble quality. First, my bias: it would imply the most social good would be if I made my own prediction market with no fees at all, or if I paid someone to do so. I don't wanna do that. Second, I can imagine two such completely-free markets (so equal EV), one of which is designed to take advantage of human psychology and the other of which is not. I think reducing to EV would call those equal when we shouldn't - so there must be some additional piece to gamble quality. I think that missing aspect is a huge part of my concern here.
3) I like the distinction between markets/contracts on the same platform. If we have a gamble quality measure, I agree with saying contracts on seahawks have lower quality than contracts on sunlight exposure in an area over time.

I feel like... there's some tie here to paying for answers that Kalshi and Polymarket are missing, but manifold gets because you can add liquidity to a market, and make your own markets.

Will have to come back and think more later!

@DannyqnOht So... let's do some thinking in metaphors.

I can offer a new flavor of ice cream to the market.

From an agent perspective, this strictly increases utility. Agents could always act as if the option wasn't there and get just as much utility as before.

As humans, we are not immune to the decoy effect. So in reality, offering a third flavor of ice cream to us non-perfect agents, in some circumstances, predictably lowers our utility. So much so that there are whole industries dedicated to pricing things to profit from these kinds of mistakes.

Is there a way we can quantify the... goodness? of a offering a new flavor of ice cream despite this? To help a hypothetical overly-moral ice cream maker make the decision of what to sell, if they aren't just deciding on pure profit?

Hopefully our measure of goodness would work as intended for both offering new ice cream flavors, and offering new gambles.

In my previous sentence about something manifold has but Kalshi/Polymarket don't, I was thinking along the lines of: "If we could measure willingness-to-pay-for-an-accurate-answer, that might be a way to measure goodness of a bet. And peoples willingness to add liquidity to a market is a pretty direct measure of that."

But that actually still has lots of flaws as a measure of goodness. Tragedy of the commons (anyone can profit from public info caused by an altruistic person's liquidity provision) means we'll underestimate how good a bet was. And beyond that, people are actually somewhat consistently bad at reacting correctly to the decoy effect - many people are bad at spotting when they should ask for 3 choices rather than 8 (no source.). So I don't expect willingness-to-pay-for access to a gamble, to necessarily correlate with goodness for society as a result of access to a gamble.

Hmmm... I mean, there's the tautology, that a gamble is good iff the net utility after offering the gamble is more than the net utility before.... But how can we measure a proxy.... And how can we quantify the amount of positive influence that online instant-access gambles (eg manifold) might have?

Well, there's a clear need for knowledge to operate in the real world. Maybe the more that need is satisfied, the better prediction markets are doing? Especially for things where google isn't working well. If suddenly no one had to figure stuff out any more, everyone just created/checked the markets; I'd say they were pretty valuable.

And the only real cost I can see here is the cost to humans with biases (which is all of us, once the machines really get to trading on these). Gambling addiction included in these biases, and called out because it's the biggest, most obvious cost. Maybe I need to define that cost more clearly.... What even is gambling addiction, and what's its cost?

I think I would classify it as... not being able to stop when you know you're going to lose. Not knowing how to walk away.

So... Is an influx of easily available, fair (relative to past ones), online gambles going to cause more of that? And how do we fit that into the 'cost' of an individual gamble...

I get the feeling (based on vibes and what I think casinos/scratchoffs do) that there's a pattern of a few small wins that optimally addicts humans to maximize the losses they end up with. Do I think modern prediction markets, intentionally or unintentionally, hit this pattern? That seems like a good guess about whether they're likely to cause hurt.

hmmm... yeah, I think they do. I think the actual variance in real world outcomes is likely pretty close to this pattern.

Wait, that doesn't make sense. Our gambling addiction patterns evolved into us. We are biased to make the kinds of guesses that happened to work in the real world. So maybe tying it to real world things can't, in expectation, select for patterns that cause losses in humans?

We'll call that slight evidence that the gambling argument isn't really important. I think it's easy to ruin that evidence, if the markets are being selected for (eg, Kalshi & Polymarket could easily select for only making markets that will resolve in ways that addict humans) but it is slight evidence in the direction of 'nah it's fine'.

Okay, shoot. I think that might answer the question.

If you're someone with control over market creation, and you care about this, your markets won't happen to be the ones that addict humans. You'll have markets like weather and stock prices and housing prices, and you won't try to pick things that are.... the things Kalshi and Polymarket pick. Sports. You wouldn't expand into that territory.

I guess this means I really don't trust the creators of polymarket/kalshi. Maybe that's all it comes down to.

The more I think about this, the more I think I should rewire my program to just let users manually put in bets, outcomes, and so on; and it'll still do the math for you. Gut the kalshi connection stuff. If people want to manually input the probabilities & bet from there, fine; if people want to manually make connectors that scrape from kalshi/polymarket, fine. But maybe I shouldn't do that for them. Only make the connector for Manifold.

I was trying to just bet on important things, not on the sports stuff. But maybe that's really not good enough if the people controlling market generation are so willing to take advantage of the markets that are making them the most money.

@DannyqnOht Ooo, wait! Idea!

I was thinking about messaging Kalshi support and saying something like "can you guys at least like... add a flag for "this market helps the world" or something? To demonstrate that you're not just trying to profit from sports betting?

But I don't need them to do that! I can do that myself!

I could keep the kalshi connection stuff. Just have a list of 'marketsDannyApprovesOf' in the code. If someone's betting on an un-approved market, they get a warning and can continue to use it - I'm not gonna stop you from doing it if you really want to.

That actually seems both good for the world, and makes me money, and makes other people money!

Even better if I'm not the one who decides what makes a market good or not....

Yeah, maybe that's a better option here. Offload the evaluation of "is this a good gamble to offer" onto someone else. Make my code actually consider it - not in a restrictive way, but in a warning-flashing-lights way.

Any ideas about how to do that?

@DannyqnOht I'm trying to keep up with your train of thought, but it's starting to take effort to keep up with that text wall.

Quick test. Can you show me the seahorse emoji? I need some reassurance that you're a human and not a reasoning model.

Link to where this started