New ask Hacker News story: Tell HN: Pangram is easily-defeatable with Claude
Tell HN: Pangram is easily-defeatable with Claude 2 by nunez | 1 comments on Hacker News. I use Pangram all of the time to detect whether what I'm reading is fully AI-generated or not. Today, I wondered how easy it was to defeat it. All I had to do was show them that the text they generated was detected as 100% AI generated to get them to generate a "human-sounding" text snippet. Claude Sonnet: https://ift.tt/0zXITy6 Interestingly, ChatGPT 5.4 won't do it, at least not with the default model it uses: https://ift.tt/2JPaMrZ. I'm guessing it can be jailbroken to do it though.