Will a single model achieve superhuman performance on all Atari environments by 2025?
Plus
14
Ṁ8982026
22%
chance
1D
1W
1M
ALL
Note: single *model*, not single *algorithm* (the latter has already been done). If I decide the model is actually many environment specific models stuck together (e.g. a hierarchical model that predicts current environment and delegates to a subagent) it will not count (a single model that *learns* a hierarchy without any inductive bias pushing it towards that *would* count, however).
This question is managed and resolved by Manifold.
Get
1,000and
3.00
1,0003.00Related questions
Related questions
Will an AI achieve >85% performance on the FrontierMath benchmark before 2028?
70% chance
Will any AI reach 20%+ performance on FrontierMath by December 31st 2026?
97% chance
Will an AI achieve >85% performance on the FrontierMath benchmark before 2027?
70% chance
Will an AI achieve >30% performance on the FrontierMath benchmark before 2026?
89% chance
Will a single model achieve superhuman performance on all OpenAI gym environments by 2025?
25% chance
By 2026 will any RL agent with learned causal models of its environment achieve superhuman performance on >=10 Atari environments?
81% chance
Will anyone prompt a GPT model to play "Colossal Cave Adventure" well by the end of 2024?
49% chance
Will any transfer learning model, trained for any amount of time on one Atari environment, outperform the median human learning curve on most other Atari environments when transferred by 2026?
45% chance
Will any model get above human level on the Simple Bench benchmark before September 1st, 2025.
55% chance
Will models be able to do the work of an AI researcher/engineer before 2027?
35% chance