Ported over from previous thread
Recently made the discovery of a website called AI 2027 tracker which grades the predictions of AI 2027 independent of the authors.
The way they grade these predictions is odd, 80% accuracy is for some reason "confirmed accurate" and "partially accurate" which is bizarre.
The website also shows a consistent degradation of accuracy over time which is funny
EDIT:ok so they do have a grading criteria, 100% is "highly accurate/confirmed accurate" and 80% is supposed to be "largely accurate". For some reason they don't include this on the timeline grading? it's either confirmed or partially accurate, so it's still weird