Resolution Criteria
YES if conditions 1 and 2 are met before market close:
Autonomous Self-Improvement (Technical Threshold): An AI system independently designs, trains, and deploys a novel successor model that demonstrably outperforms itself across major reasoning and coding benchmarks without human algorithmic intervention. And that new child can and does create one better than itself again. (I'm sure there will be official tests for whether it is truly better)
Consensus & Verification (Validation Threshold): The achievement is confirmed by at least two of the following:
Government / Safety Verification: An official statement from a major AI Safety Institute (e.g., US AISI, UK AISI).
Expert Consensus: A majority consensus in a formal survey of top AI researchers (e.g., NeurIPS/ICML/ICLR authors).
Mainstream Media: Front-page/lead headline coverage by at least three major news outlets (NYT, WSJ, BBC, Reuters, FT).
NO
if the above conditions are not met within an option's specified timeline
Unverified corporate marketing, leaks, or routine prompt/fine-tuning updates do not qualify.
For this market, the speed at which the model improves doesn't matter; "the singularity" occurs once the essential training process of a child AI has begun and is entirely devoid of humans. (aside from obvious human maintenance of servers etc.)
Note: the human-unrestrained AI intelligence explosion is what this market focuses on. Not whether the AI goes the extra step to begin wreaking havoc everywhere.
BACKGROUND
The Singularity describes the moment machines no longer need humans to keep getting smarter.
“We’ll be in a state of perpetual future shock for a number of years probably.”
- Nat Friedman (co-leader of Meta Superintelligence Labs)
The part above is an excerpt of this article: https://colossus.com/article/sarah-guo-conviction/ (~part related to this market ~ 3 minute read)
1,000People are also trading
@ChristopherRandles no, the market focuses on intelligence explosion. althoguh just and interesting side note: unless AI is extremely heavily restricted, once it can increase its own intelligence so that it can better increase its intelligence once again, maybe it will get rid of humans if put in the wrong hands?
check this one out too:
https://manifold.markets/post/searching-for-quality-markets-and-f?r=Q2hpbXBleQ
@Stralor yep 😁i think this market is great. i put lots of thought into this, but it worth it i think 😁
@Chimpey Unfortunately, for this type of market, the trader bonus only triggers once, even if they trade on all three.
This is a very strange definition of "the singularity." Leaving that aside, the criteria also make no sense: an AI being able to design a new AI independently, won't save it from a human unplugging it. "The world" won't be able to "no longer need humans" until a machine can keep itself powered, repair itself, etc., without human intervention.
@ChurlishGambit hmm... lemme just delete that part then. It was just from an article I thought was kinda cool
@Chimpey to your other comment, it says "no longer need humans to keep getting smarter" not just "no longer need humans"
@ChurlishGambit also, at some point, which is not related to the market very much, but at some point the AI will prob be able to leave to other servers an copy itself and whatever happens happens etc. a quite flawed but fun example below: What You'd Actually See During an AI World War
