I don't use AI to write any of my text mostly because (a) I think I'm a better writer, and (b) the AI can't draw on my experience the way I can. The first may be hubris and is entirely subject to change, but I can't imagine the second ever changing. I don't have any opposition to AI, and I use it to generate code and to remove noise from my photos. And it's OK with me if other people use AI to write (though I really wish they would optimize for brevity). But key in all this is that AI is a tool that should be subject to the wishes of the person using the tool, and that's where Anthropic's watermark fails. "Watermarking is misdirection: a way for goods to appear to do one kind of thing while secretly doing another. Real writers (e.g. Gruber, Teare, and Jarvis) are calling bullshit on it." This is really important, because if it can secretly do one thing, then it can secretly do any number of things. I can protect my code with tests and my images with my eye, but there's no way to protect my writing.
Today: Total: [Direct link] [Share]
