Context:
In the recent Dwarkesh Patel interview, Paul Christiano gave an AI capable of building a Dyson Sphere a chance of 15% by 2030 and 40% by 2040.
Is this plausible?
Question:
When we will have an AI capable of building a Dyson Sphere?
Notes:
- I will not bet in this market.
- Answers will resolve at the date is hit, or we develop AI.
- I make rulings on my markets based on the spirit of the question, not the strict rules as written.
Can you provide a little more context for what you mean by capable? Does the AI have to be capable of only devising plans for the dyson sphere, and building it if it had enough time and resources, or does it also have to be capable of leveraging enough productive capacity to actually produce a dyson sphere by the stated date?
@Stralor I've got pretty large limits on after 2050 (500 @ 85, 1000 @ 80) if anyone disagrees and wants a piece
@jgyou We are definitely not currently capable of building a Dyson Sphere. And yes Time would probably be a factor.
Recommendations on ways to operationalize this are welcome though. If you think a given date would resolve YES, maybe we could just ask whatever the AI system in question is how to resolve this market. ;)