Will we have at least one more AI winter before AGI is realized?
➕
Plus
207
Ṁ70k
9999
38%
chance

Resolves YES if general consensus is that we are in another AI winter comparable to the AI winters of the 1970s and 1990s

Resolves NO if general consensus is that we have realized AGI.

Closing date will be moved forward perpetually untill either of the two resolution critera have been met.

Get
Ṁ1,000
and
S3.00
Sort by:

Nah

maybe there will be an Ai winter between level 3 and level 4 on the OpenAi scale.

level 4 on the OpenI scale is literally ASI

once AGI is achieved by level2 and perfected by level 3 maybe there would be an Ai winter before ASI is reached

Yeah, AGI is not a thing. Literally everything until the heat death of the universe will occur before AGI is achieved.

sold Ṁ50 NO

How would you define Winter? Wikipedia says: "AI winter is a period of reduced funding and interest in artificial intelligence research". But by how much would funding & interest need to go down to count as a winter? For example if funding were to go down levels between 2015-2020 that would still be much higher than previous winters, but much lower than currently.

@JanW Whatever definition is general consensus at the time.

@komplexkonjugat My estimate really depends on the definition. I think it's likely that investment will drop at some point, but I think it's quite unlikely to go back to levels before 2018. For the same reason, I expect that people in the future will disagree about whether something counts as an AI Winter for the same reason people might.

I feel like it's possible to decide now on a good resolution criteria like "Investment in a year drops to 2018 levels of investment" or "Investment in a year drops by 50% compared to the previous 5 years"

predictedYES

@JanW Then go create a market with that definition. I won't change this market.

© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules