I was also looking at https://aiimpacts.org/2022-expert-survey-on-progress-in-ai/#Data (the dataset itself; the summary doesn’t include the milestones). This new version seems like total garbage. The experts continue to predict several of the milestones are five years out, including milestones that were achieved by ChatGPT (ie a few months after the survey) and at least one milestone that had already clearly been achieved by the time the survey was released!
I’ve only given the data a quick look and found it hard to analyse—but yeah, many of the forecasts look bad. But some of the medians (I think- from eyeballing data) seem not terrible—the ‘create top forty song’ shifted from 10 year median to ~5 year median. The ‘answer open-ended questions’ shifted from 10 year median to ~3 years.
But like you say, for many of the milestones I resolved as being met before this survey went out, they still have medians >0 years from now so—if I’m right in my judgements—the experts seem pretty poorly clued up on recent developments across the field.
I’ve only given the data a quick look and found it hard to analyse—but yeah, many of the forecasts look bad. But some of the medians (I think- from eyeballing data) seem not terrible—the ‘create top forty song’ shifted from 10 year median to ~5 year median. The ‘answer open-ended questions’ shifted from 10 year median to ~3 years.
But like you say, for many of the milestones I resolved as being met before this survey went out, they still have medians >0 years from now so—if I’m right in my judgements—the experts seem pretty poorly clued up on recent developments across the field.