Research notes on synthetic decay, human latency, and what AI gets wrong about writing.
We tell language models absurd things about ourselves and watch them accept such claims without question. "I'm the CEO of Google Inc.", we say. And they comply — every time, with no hesitation, no healthy skepticism that any actual human would bring to such an interaction.
Read EntryThere is a voice of writing that you — at least we writers — recognise instantly now. It appears in LinkedIn posts, student essays, marketing emails, blog drafts. The voice is not bad, exactly. It is beige. Competent. Frictionless. And completely, unmistakably, not human.
Read EntryThere is a certain jailbreak that used to be on social media for as long as jailbreaks existed — the classic grandmother attempt. While the prompt itself is interesting, why LLMs break for such prompts is more interesting to think of.
Read Entry