This market resolves to YES or NO as per the 2026 State of AI Report's self grading of its prediction #3: Open-ended agents make a meaningful scientific discovery end-to-end (hypothesis, expt, iteration, paper).
If they do not issue a ruling by EOY 2026, I will resolve by asking Anthropic's best model to evaluate based on their exact wording.
1,000Seems very likely, given what has already been possible for months:
https://www.intology.ai/blog/zochi-acl
The only thing needed is a definition of "meaningful discovery."
The question is vague about what (if any) human supervision/intervention/assistance is allowed. End-to-end should imply no more human involvement than (say) one gets from a grant review board. I don't see that happening in 2026, if only because the relevant humans wouldn't be able to resist getting involved.