I handed over a finished manuscript and asked AI to help me edit it. It rewrote passages and changed plot points, and it was right to — I never said what I meant by "edit." What that cost me, and what it taught me about listening to people.
A stranger stopped me ten seconds after I walked off the stage with a second place trophy and offered to tell me what I'd done wrong. He may have been right. I'll never know.
There's a finding in AI research called emergent misalignment. Researchers took an aligned language model and fine-tuned it on one narrow thing: examples of insecure computer code. Bad code, nothing else. No malice anywhere in the training data. The model didn't just start writing bad code. It started giving harmful advice on subjects that had … Continue reading Estimating
In January of this year I used an AI tool to help me edit my novel. The chapter that came back wasn't mine. Same characters, different scene. Events I hadn't written, in a book I'd already spent a hundred and five thousand words on. I got about halfway down the page before I stopped and … Continue reading What I Didn’t Say
Not long ago I got into a technical conversation with a guy at work, and I was dead certain I was right. Certain he was being short-sighted. Certain he wasn't really hearing what I was saying. He wasn't hearing me. But not for the reason I assumed. I hadn't framed my thinking well enough for … Continue reading Why Is This Not Working?!
David Willis 