
Will AGI be more attributable to bottom-up machine learning than top-down symbolic systems?
Plus
24
Ṁ8952300
86%
chance
1D
1W
1M
ALL
One of the questions from https://jacyanthis.com/big-questions.
Resolves according to my judgement of whether the criteria have been met, taking into account clarifications from @JacyAnthis, who made those predictions. (The goal is that they'd feel comfortable betting to their credance in this market, so I want the resolution criteria to match their intention.)
Clarifications:
Causal attribution in such cases is imprecise. I mean something like: If A and B both seem to be necessary conditions for C, and a 30% reduction in A is twice as likely to make C fail to occur as a 10% reduction in B, then C is twice as attributable to A as to B.
This question is managed and resolved by Manifold.
Get
1,000and
3.00
Related questions
Related questions
Are LLMs capable of reaching AGI?
62% chance
Will we consider the first AGI system to be neuro-symbolic?
Will the first AGI be built mostly within the deep learning paradigm?
83% chance
Will AGI be interpretable due to CoT and reflection and similar methods?
31% chance
Will we have an AGI as smart as a "generally educated human" by the end of 2025?
57% chance
Will AGI be made up of narrow forms of AI working together when it is first reached?
33% chance
Will the first AGI be a large language model?
30% chance
Will AI create the first AGI?
41% chance
Will the first AGI be an LLM that emulates Nobel-prize-worthy scientific research?
26% chance
Will humans create AGI, either directly or indirectly, within the next 24 months?
16% chance