It's still a great idea - I haven't heard anyone else suggest using such context for predicting AGI timelines.