[Manifold May Showcase] Will 25+ users nominate markets?
[Manifold May Showcase] Will 25+ users nominate markets?
56
1.1kṀ52k
resolved May 29
Resolved
YES

This market is where non-Trustworthyish users may propose one market to showcase, as outlined in the Showcase group rules:

  • Every user may leave 1 top-level comment per round

  • Top level comments must include a link to the proposed market and a description of why the market merits inclusion

  • Markets that 3+ Trustworthyish users "like" will be added to the Showcase by the Manifold team

    • This means they will receive M10k liquidity subsidy and be eligible for the end-of-round bonus

This market itself will resolve YES if we get 25 or more good-faith comment proposals (excluding duplicates or anything else I judge not to count.)

Get
Ṁ1,000
to start trading!

🏅 Top traders

#NameTotal profit
1Ṁ337
2Ṁ104
3Ṁ101
4Ṁ89
5Ṁ70


Sort by:
1y

Not sure where the June nominations should go?

predictedYES 1y

@Austin I believe the resolution criteria have been satisfied 🙂

@Austin maybe it's after the time limit, but I'd still like to nominate this market. It is clearly a milestone, a big news event, a topic of interest to people from EA, AI, AI safety communities for sure.

Out of the 72 traders so far it's got under 10 predicting YES and everyone else buying NO. It has potential to demonstrate how insider+public information aggregation can happen nicely through prediction markets. It has a well defined resolution criteria but still demonstrates how mispricings are corrected in short term markets


I would like to nominate this market, because:

  • The market considers an economically, politically and strategically important subject.

  • Has tangable value for planning individual actions into the future, as well as the future and prosperity of the region.

  • Has an abstract and clear resolution criteria.

1y

I'd like to nominate this market:

Algebraic value editing works better for larger language models, all else equal50%

This is interesting to me because it's a market about research that was featured in that research. Often after research is published people tend to dismiss the results as "that's obvious" and I think prediction markets can be a good way of preregistering outsider claims about obviousness.

I'm going to subsidize it myself to put my mana where my mouth is, but I'm also nominating it here.

https://www.lesswrong.com/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector?commentId=gqxtrXH88HYc8gYFn

1y

I would like to nominate this market:

The market is clearly an important advancement tracking market, and has clear resolution criteria, and the creator has been very proactive in clarifying details about what counts. There are 100+ users trading in this market, which means it's a popular topic, and will be a major milestone in progress of AI if this is achieved.

1y

This market has not really crossed 50% but if it does, then it will be a lot of surprise to me, at least, and presumably to other users as well!

predictedYES 1y

Can we get a current count of nominated markets? I count 25 now but not sure which ones count and which don’t.

predictedYES 1y

I nominate this market, as it has potential of tracking Russia's global intentions. Imagine, that you could see where the Sauron Eye is looking.

I nominate this market (part of a conditional pair). Next year, this will be directly relevant to most American longtermists' voting decision. Even if the pair remains within-the-noise of each other, that would still be useful for cause prioritization.

1y

@ScroogeMcDuck I don't think either of those markets provide any meaningful information. The only account that bought yes on either question is a bot that didn't read the question, and the decision to bet no is based more on non-specific factors like interest rates and opportunity cost than the questions themselves at this point

predictedYES 1y

@ScroogeMcDuck I certainly hope markets like these are not directly relevant to any longtermists' voting decisions. The probability of AI killing 10% of the human population by 2028 is so low that markets like this don't actually provide any meaningful insight into them (And that's not even accounting for the "Don't bet YES on an event so catastrophic that you won't care about your mana anymore if it happens" effect). Markets can only be well-calibrated for events that have a reasonable chance of occurring.

Even if the market was well-calibrated, long-termists would be better off prioritizing things that are of greater risk/importance than "AI wipes out 10% of humanity by 2028". There are other existential risks (e.g., nuclear war, bioweapons, AI killing us at a much later date) that have a far greater chance of occurring.

1y

@JosephNoonan I agree it's too low for the markets to do a good job, since the markets do poorly at distinguishing any probabilities below a couple percent is. But to put a number on just how low or high the probability is: https://www.metaculus.com/notebooks/8736/a-global-catastrophe-this-century forecasts 18% chance of a "global catastrophe" (loss of at least 10% of the human population) this century. And it puts AI as the most likely cause, slightly ahead of synthetic or natural pandemics and nuclear war. (Note - this is talking about global catastrophe; if we look at extinction events, AI is forecasted by to be far more likely the source). So it's not extremely low, and questions like this are worth considering, markets just aren't the best way to do it.

predictedYES 1y

@jack But the market in question is only asking about an AI-caused catastrophe before 2028. That probability is extremely low, since, for it to happen, we would have to develop an AI capable of causing such a large catastrophe in a short time span. By any reasonable estimation, we are nowhere near an AI that capable, so I don't consider an AI catastrophe on this scale happening before 2028 to be even remotely plausible. It gets much more plausible as the time window increases, since then there is more time for a dangerous AI to be developed.

1y

@JosephNoonan I think it's more likely than you think. https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/ has a forecast of 50% by 2032 for AGI (high competency on diverse fields of expertise, including coding and robotic capabilities), which I think is a quite reasonable forecast. It's roughly 35% by 2028. Forecasters expect AI capabilities to develop very quickly, a lot of the probability mass of an AI catastrophe is sooner than many people expect.

predictedYES 1y

@jack Even if that estimate is correct, an AGI =/= "AI capable of wiping out 10% of humanity". I still find in extremely unlikely that, if an AGI is developed by 2028, it will also wipe out 10% of humanity in that time frame.

1y

@JosephNoonan But actually among longtermists, there is disagrement about 4 year risk. I agree it's lower than certain prominent people think it is, but that doesn't mean this isn't informative.

1y

@PatMyron To be fair, we are a long ways from the election these are recent.

1y

@JosephNoonan I'm very well aware that there are more steps to get to AI that wipes out 10% of humanity, but I would forecast that in worlds where AGI is developed in the next 5-10 years, there is also a significant probability mass on AGI causing a global catastrophe in a very short timeframe as well. The chain-of-reasoning here features steps that are highly correlated - short timeframes for step 1 are highly correlated to short timeframes for the following steps. Even if it's a 1% chance in the next 5-10 years, that's about as big as other global catastrophic threats like nuclear/pandemic/etc in the same timeframe.

1y
1y

@MarcusAbramovitch Without having thought too much about it I would pick the soonest close date (2028)

I am suggesting my market on traffic regulations in Paris, which has the potential to spark a crucial conversation regarding this pressing issue in other European cities. This policy aims to enhance the quality of life in our urban centers, which are currently dominated by cars, but also touches on climate change and the future of individual transportation. Despite growing interest among people to ban cars from their cities, such proposals often encounter resistance. While Paris has declared its intention to implement this policy, there remains some uncertainty surrounding its implementation.

The market asks about a specific policy, has clear resolution criteria and is relatively short-term. It is also important for me personally since I am considering moving to Paris next year :)

1y

I'd like to nominate this market

https://manifold.markets/Nostradamnedus/will-there-have-been-a-noticeable-s

While the resolution criteria are a little vague and it will be a subjective judgement call as is, the market has done an impressive job soliciting meaningful comments on topic and I anticipate it will continue to do so.

There is value in precise predictions, but there is also value in markets surfacing relevant information.

What is this?

What is Manifold?
Manifold is the world's largest social prediction market.
Get accurate real-time odds on politics, tech, sports, and more.
Or create your own play-money betting market on any question you care about.
Are our predictions accurate?
Yes! Manifold is very well calibrated, with forecasts on average within 4 percentage points of the true probability. Our probabilities are created by users buying and selling shares of a market.
In the 2022 US midterm elections, we outperformed all other prediction market platforms and were in line with FiveThirtyEight’s performance. Many people who don't like betting still use Manifold to get reliable news.
ṀWhy use play money?
Mana (Ṁ) is the play-money currency used to bet on Manifold. It cannot be converted to cash. All users start with Ṁ1,000 for free.
Play money means it's much easier for anyone anywhere in the world to get started and try out forecasting without any risk. It also means there's more freedom to create and bet on any type of question.
© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules