SoAI 23 3/10: Will Self-improving Al agents crush SOTA in a complex environment (e.g. AAA game, tool use, science)?

33

1kṀ8977

resolved Oct 11

Resolved

NO

1H

6H

1D

1W

1M

ALL

In the 2023 State of AI Report (https://www.stateof.ai/) this is the third prediction:

"Self-improving Al agents crush SOTA in a complex environment (e.g. AAA game, tool use, science)."

This resolves to YES if the 2024 State of AI Report says this prediction was true, and the example they site is reasonable.

This resolves to NO if the 2024 State of AI Report says this prediction was false, or all cited examples and ones I can come up with seem unreasonable.

I will resolve this early if the market is >95% and I believe we have a clear example of such an advancement, but since we do not know when the report comes out we cannot resolve to NO early.

If there is no report by EOY 24 I will evaluate this myself.

When evaluating advances, I will interpret the word 'crush' to mean a large advancement in capabilities or skill - e.g. merely 'the new best chess or go program' would not count but AlphaZero or AlphaGo would have.

State of AI Report 2023

The State of AI Report seeks to trigger informed conversation about the state of AI and its implication for the future.

Technical AI Timelines

Get

1,000

to start trading!

🏅 Top traders

#	Name	Total profit
1		Ṁ612
2		Ṁ438
3		Ṁ385
4		Ṁ299
5		Ṁ282

People are also trading

Will computer-using AI agents play Balatro better than me by eoy 2025?

By 2026 will any RL agent with learned causal models of its environment achieve superhuman performance on >=10 Atari environments?

Will AI be Recursively Self Improving by mid 2026?

Will an autonomous agent resolve 90% of tasks on SWE-bench by 2026?

Will the ARC Prize Foundation succeed at making a new benchmark that is easy for humans but still hard for the best AIs?

When will AIs be good at solving complex problems? (read description)

Will AI be capable of fast self improvement before physical AI robots are massively used for improving AI capabilities?

In 2028, will an AI be able to play randomly-selected computer games at human level, given the chance to train via self-play?

Will any AI solve more than four of AI 2027 Marcus-Brundage tasks in 2025?

Will AI top level capabilities generally be judged by question and answer benchmarks in 2029?

Related questions

Will computer-using AI agents play Balatro better than me by eoy 2025?

By 2026 will any RL agent with learned causal models of its environment achieve superhuman performance on >=10 Atari environments?

Will AI be Recursively Self Improving by mid 2026?

Will an autonomous agent resolve 90% of tasks on SWE-bench by 2026?

Will the ARC Prize Foundation succeed at making a new benchmark that is easy for humans but still hard for the best AIs?

When will AIs be good at solving complex problems? (read description)

Will AI be capable of fast self improvement before physical AI robots are massively used for improving AI capabilities?

In 2028, will an AI be able to play randomly-selected computer games at human level, given the chance to train via self-play?

Will any AI solve more than four of AI 2027 Marcus-Brundage tasks in 2025?

Will AI top level capabilities generally be judged by question and answer benchmarks in 2029?

© Manifold Markets, Inc.•Terms•Privacy