Will I think that alignment is no longer "preparadigmatic" by the start of 2026?
➕
Plus
51
Ṁ9837
2026
41%
chance

Many people in alignment believe that alignment as a field is currently preparadigmatic. There are a number of competing paradigms with different supporters, without much consensus on a fieldwide paradigm.

Some examples of when this question resolves as Yes:

  • Most alignment researchers that I know agree that a particular paradigm (or small family of paradigms) is the best frame for alignment and captures the "hard part"

  • Most alignment researchers that I know fall into one of a small number (<=3) of camps supporting different paradigms, but most alignment researchers that I know agree that all of these paradigms capture the "hard part" of alignment (but disagree which frame is the most "natural" one).

Some examples of No resolutions:

  • There continue to be a large number of different paradigms.

  • There is substantial disagreement over which paradigms capture the "hard part" of alignment, or what the "hard part" even looks like.

This market is pretty subjective.

Get
Ṁ1,000
and
S3.00
Sort by:
opened a Ṁ750 YES at 50% order

Limit orders up

Depends, how did you update on the recent Anthropic paper? https://www.anthropic.com/news/mapping-mind-language-model

bought Ṁ10 NO

As organizations grow, but continue not trying to resolve their differences, we may be moving towards more established paradigms that each understand the others even less than they do now

@TaoLin If there are a small number of major paradigms that don't understand each other, presumably this falls under example #2 under NO.

Would "pluralism" be considered a paradigm?

I think this is largely a social question rather than a technical one, and I don't really see a route to "consensus" without extremely strong technical results/demonstrations, which I don't think are that plausible.

© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules