10h agoPolicy scholar Dean W. Ball argues regulators may classify AI benchmarks to hide their inability to tighten the 10^26 FLOP thresholdRegulators face uncertainty over self-reporting versus independent model testing.SentimentSentimentPos0%Neg100%Users reacted negatively to a policy analyst's prediction that regulators will classify AI benchmarking to conceal weak thresholds, calling it the beginning of the end of the good timeline due to concerns over private benchmarks.1 comment with sentiment. View comments.