[Question] Modeling AI milestones to adjust AGI arrival estimates?

Is there a roadmap of ma­jor mile­stones AI re­searchers have to achieve through be­fore AGI is here? I feel like this would be benefi­cial to up­date on AGI ar­rival timelines and ad­just ac­tions ac­cord­ingly. Lately, a lot of things that can be clas­sified as a mile­stone have been achieved—GLUE bench­mark, StarCraft II, DOTA 2 etc. Should we ad­just our es­ti­mates on AGI ar­rival based on those or are they mov­ing ac­cord­ing to plan? It would be cool to have a place with all fore­casts on AGI in one place.

No comments.