🏅 Top traders

It's past 9AM PST 1 Jan 2024 and to the best of publicly available info, OpenAI's board of directors consists of Bret Taylor, Larry Summers, and Adam D’Angelo (as listed on OpenAI's Our structure page). Resolved YES to those three individuals and NO to everyone else.

@cash when are you planning to resolve this market?

@KevinBurke "This question resolves YES to all the named individuals who are on the board of directors of OpenAI the Monday morning 1 Jan 2024 (9am PST)"

@cash Right sorry just you could be like "I will wait in case they voted on a change but didn't actually announce it publicly" or whatever. I think that is the right policy

@KevinBurke As you pointed to earlier, OpenAI has updated their Our structure page to three named individuals. So long as nothing changes before 9AM PST tomorrow, I'm inclined to see that as sufficient to conclude that the board consists, as of that time, of those people. Even if a new board member were announced later tomorrow, it might not be clear that this is announcing something they did in the past, as opposed to something that will happen in the future (clarity might not be possible, for example, it's not clear if Emmett Shear was ever technically CEO). Feel free to let me know your rationale for waiting.

Maybe they will announce some new board members over the long weekend

Three business days left for OpenAI to announce new board members or for Bret Taylor to have a change of heart (and announce it)

Clock is ticking for OpenAI to announce some new board members

It seems like the combined probabilities of all the non-top-3 options (the current board) implies something like a 80% chance that OpenAI adds at least one board member before January 1. I'm guessing most people voting up any given individual don't actually believe there's an 80+% chance that OpenAI adds a board member in general before Jan 1, but do believe in their given candidate being added above the current odds, and are too lazy/illiquid to go about buying no stakes on everyone else — which would seem like a prediction market failure.

Does this have a name / has anyone else discussed this phenomenon?

EDIT: Created a new market to test this. If it is the case that there's a big gap between the prediction from my market and the implicit prediction from this market, I'll assume this is an inefficiency in extrapolating implicit probabilities from multiple-choice markets. If not, here's some free arbitrage for anyone who wants to correct this market or my own.

@TylerJohnston This is due to poor liquidity because each board member is an individual yes/no market.

Individual yes/no markets like this generally have "deceptive" liquidity where some options can be way less liquid than others.

@TylerJohnston also it could be an 80% chance of adding one member, or a 40% chance of adding two, or a ~30% chance of adding 3 to bring the board size back to the original 6.

@RyanGreenblatt Interesting — so a bunch of small individual markets here are overpriced due to low trade volume/liquidity, and there's no incentive to correct them because of the fact that they are low-liquidity + the correction is pretty small?

I wonder if there could be some way to design a platform that lets me trade the conjunction of two or more markets resolving yes/no. Some sort of design that incentivizes people to correct a bunch of small low-liquidity markets at once. Like, what if I could place a bet on all of the non-current-board choices resolving "NO," in such a way that I get favorable odds (80% in this case) and it actually impacts the current odds of the individual markets rather than requiring a single counterparty for the specific bet (which would basically mean creating a new market). I don't imagine that'd be easy or even possible to design though, especially since these aren't independent events.

OpenAI website indicates Bret Taylor, Larry Summers, Adam D’Angelo are currently board members. https://openai.com/our-structure

can someone who is betting NO on eg Bret Taylor explain their rationale? they might not even have a board meeting in the next month.

@KevinBurke I've made a decent amount of mana going onto all of the OpenAI markets and betting a small amount of YES on any halfway plausible option under 5% and NO on any halfway plausible option over 95%.

I expect that I'll probably lose these particular bets, but I still think the chance that e.g. a board member leaves within the first month is probably at least 10%.

@KevinBurke In addition to what @FaulSname said, the percentages in this market have just been way too high. Nonprofit bards are often just three members, and while it could go up to five or seven, that usually takes time. I just bet NO on all the options above ~10% (approximately the current risk-free rate on Manifold) and have made a decent profit as well.

@Jacy I'm specifically talking about Bret Taylor who was named as a board member in a press release put out yesterday by OpenAI.

@KevinBurke I'm personally selling when it's around 95% because of the opportunity cost of that money sitting there for a month

@KevinBurke I'm NO on all because OpenAI said that there will be more changes to the board. That might happen in december.

@Shump To be clear are you NO on Bret Taylor?

@KevinBurke I'm NO on all current board members, including him

Do non voting members count?

@Odoacre Yes they're members of the board. Technically called non-voting observer members, but members nonetheless

@Odoacre Good question. I'm leaning yes. The original description is worded as who are the "members" (and the question as who is "on" the board) and not who are the "directors". From what I am reading, non-voting observers are not "full" or "formal" members by some sources, but that still by implication makes them "members" of the board. Some news sources are referring to Microsoft's non-voting observer as a board "seat", which sounds like membership. From what I understand, non-voting observers are codified by the organization's bylaws and have confidentiality obligations, which sounds like attributes of membership. In any case, the presence and identity of a Microsoft non-voting observer in board discussions is relevant to how OpenAI's board will function in-practice, and so seems worth tracking as part of a question about "who is on the board" of OpenAI. Let me know if anyone has relevant arguments I should consider, otherwise, I'll edit this qualification into the question description tomorrow. If we don't learn the identity of the non-voting observer for whatever reason, then I will ignore them when the resolution date comes along (I won't resolve to e.g. "Microsoft person" or whatever).

https://www.theinformation.com/articles/openai-isnt-expected-to-offer-microsoft-other-investors-a-board-seat