Sol's second cluster addressed cross-era consistency: E2-168 should use named BEA investment concepts with double-count prevention — E3-257/E4-319 need scope hierarchy so cumulative $5T and annual $1T are comparable — E3-261/E5-357 need frozen classifier and snapshot rule matching the S&P basket fix — E2-184 (30% at one or more labs) flagged as potentially easy via Google's TPU dominance, requiring either promotion or aggregate frontier-lab constraint