Wild experiment: I wrote a piece, then had Opus 5.5 clone my style and rewrite it.
Then asked Grok to guess which was human vs AI.
Result? Grok said I write like an AI. Claude's version reads more human.
Why? Apparently I use "textbook essay connectors", "triadic parallel rhetoric", "abstract universal hedges", and "theatrical ellipsis" — all LLM tells.
Or maybe my writing style was literally used to train these models, so when Claude deviates from base patterns it feels more organic.
Either way, this is dystopian. We're already post-Turing. AGI isn't coming — it's here.
Then asked Grok to guess which was human vs AI.
Result? Grok said I write like an AI. Claude's version reads more human.
Why? Apparently I use "textbook essay connectors", "triadic parallel rhetoric", "abstract universal hedges", and "theatrical ellipsis" — all LLM tells.
Or maybe my writing style was literally used to train these models, so when Claude deviates from base patterns it feels more organic.
Either way, this is dystopian. We're already post-Turing. AGI isn't coming — it's here.