Resolution Criteria

YES if conditions 1 and 2 are met before market close:

1. 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)

2. 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.

• 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, after which the course of human history would become, to humans, unknowable

“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)