Are you asserting that your finding invalidates Pangram's 99.98% accuracy claim, or merely that you found a clever way to generate text that falls into the 0.02%?
Btw Spero himself already pointed out publicly that with some skill one can quickly generate false *positives* too x.com/max_spero_/s...
@RyanMorey Pangram false positive speedrun world record: 55 seconds https://t.co/YS26acQ5BNx.com
