🏅 Top traders

@Flowers I think this was a good market and I think it'd be neat if a new one was created for the next 2 weeks starting today. Feel free to make it, but if you don't want to lmk and I might make it since it seems worthwhile imo

Aready exists! https://manifold.markets/SCS/will-openai-begin-rolling-out-the-n?r=U0NT

oh, epic

I don't fully understand why you were betting this so high even on the last day, but I really respect how well you handled the resolution despite the personal cost, Flowers.

Now that the market is over - and I must say the back and forth has made it one of the more memorable markets I've participated in - can any of the last-minute YES bettors explain your logic to me? @Peter1169, @MochiScreenTime, @Mira? Were you guys hoping this was some kind of elaborate con by Flowers, or do you just think any given day has a fairly high chance of OpenAI releasing a specific product, or what?

I was forwarding an information I got to my twitter followers, that led to today. I was very confident in the information, but apparently it’s just a very closed alpha access to only a few people.

Anyways. I made a new poll :D

Feel free to play again with me.

https://manifold.markets/Flowers/how-many-weeks-until-openai-is-roll

After learning that there was an update to iOS that included the UI for the new Voice mode, I thought the chance of the rollout starting today was big enough to risk losing 1000M in exchange for the chance of gaining ~4000-7000M if YES when the market was at 3%-1%.

In the end, the rollout didn't start today and I even managed to get 147M in benefits by selling the shares beforehand. But I still believe that losing 1000M (even if my liquid mana is usually around ~400 or less) for that chance was worth it.

So my logic for my YES trades was basically a high-risk bet for high benefits that I was willing to make.

@Flowers Ah, that explains a lot. I guess I wasn't totally off-base with "super secret insider tip".

If I'd known, I wouldn't have been so confident! Evaluating leaks is hard, and it sounds like yours wasn't even fake.

@Flowers, just to be clear, if it doesn't ship within the next 10 hours this resolves NO, right?

Yes

Do you have some sort of super secret insider tip telling you that's going to happen, by any chance?

No

I wanted to bet but as if it was two weeks starting now. So I've made a new version of this market, but with a cutoff date of July 15th.

https://manifold.markets/SCS/will-openai-begin-rolling-out-the-n?r=U0NT

@MugaSofer Bro is my fiercest opponent

@MugaSofer Update: He will outdo me just before the end 💀

Hey, you're still ahead on the Holders leaderboard thing.

You got me to dip into my prize point "savings" for this final push (and a couple of other things resolving soon I just couldn't let go.) If they really do end up releasing it tomorrow, you fully deserve it.

Oh snap, a sudden @Mira appears! I can't match the two of you combined!

So, @Flowers I think you'll find that questions like these seem pretty concrete and objective to resolve but can often be extremely challenging. You can look at the market from a few weeks ago about WWDC for instance: https://manifold.markets/Predict/will-apple-announce-a-partnership-w-db8a0fd6ca2f

I think you mean well, but in the future I'd recommend that you avoid betting in your own markets unless they're EXTREMELY objective to resolve (like, "Will Michael Phelps win a gold medal in the 100 freestyle at the 2024 Paris Olympics" or something similarly unlikely to result in any confusion or subjectivity). I hope that this one has a cut-and-dry answer, but there are a TON of edge cases:

-What if they roll it out but only selectively to a subset of users?

-What if they roll it out but it's markedly different from what they showed in the presentation, perhaps in a very limited capacity?

-What if they ship to users but only through a third party?

-What if they launch a new model called GPT-4.5 and roll out the voice model there, not within 4o?

-What if it's launched exactly around the end time of the market and it's complicated to adjudicate whether it was launched before or after since some people claim they don't have access to it then while others do?

You can see how it's quite easy to brainstorm a bunch of resolutions that might piss off some bettors on the question while pleasing others. If you're heavily invested (indeed basically all your net worth is invested in this question) it's hard to imagine that you're completely objective about the resolution!

To avoid unpleasant situations such as this, and even the appearance of impropriety, it's best to avoid betting in your own markets, especially this heavily!

Yes, could have specified better. Look at my other bet https://manifold.markets/Flowers/will-2024-see-a-major-viral-malicio

There, I included a glossary of every word of the question to get rid of any ambiguity.

However, I won’t try to find a reason to resolve yes, I won’t stretch out for the mana. If OpenAI does not begin to give access to random users, Monday at the latest, of the new voice mode, as shown in the keynote, be it called 4o or 4.5, I will resolve no. Don’t worry.

There's a cost to not letting creators bet on their own markets too, and intuitively Im not willing to pay it for all but the EXTREMELY objective to resolve

What is that cost Roberto?

@Flowers I think the problem on that market is that “major” is extremely subjective.

It dissuades people from creating markets since it's both more boring and less profitable for them. And, if created, you lose their input - in either the form of valuable information content or free dumb money subsidizing the market - and they typically are one of the people that most care about the topic.

In huge and contentious polarizing markets Im with you but thats a minority

Does anyone know how reliable manifold markets tend to be where the largest trader is also the question author at the same time, betting in their own markets? @Flowers, the question creator, spent Ṁ3,351 on YES bets to date. Runner up is @MugaSofer with about Ṁ2,626 in NO. Are there norms against this or recourse mechanisms to help if the resolution is clearly incorrect, or even just questionable?

Why wouldnt you be allowed to trade your own bets? This would only be questionable if the person already knows the outcome beforehand.

I clearly defined how it will be resolved. There is hardly any ambiguity here.

Maybe it's perfectly fine, I'm trying to understand if this is considered concerning on this website or if it's entirely within the norm. I don't mean to imply foregone conclusions. I haven't been betting for long and I've seen on another market comments discussing a somewhat similar situation after which the author fully divested IIRC. Trying to seek more info on what's good around this.

I wouldn't worry, the mods would intervene if the creator tried to blatantly resolve wrongly.

There's some potential for conflicts of interest in ambiguous resolutions, especially when the question is inherently subjective, but this seems pretty unlikely to be ambiguous; either the feature ships or it doesn't, right?

Edit: That said, I did lose a chunk of mana over the edge case where GPT-4o wasn't clearly a "GPT-4.5" equivalent or not. I don't think that was a conflict of interest issue, though, that was just genuinely an unanticipated edge case where reasonable people disagreed.

What if it only partially ships? I guess that's still a YES, right? Even if a very limited beta cohort is invited, i.e. 0.1% of their userbase? cc @MugaSofer @Flowers

"To the first users", so yeah, a partial rollout would count.