Post

It’s unacceptable for a tool to sacrifice an iota of clarity, coherence, meaning, quality, etc. for the purpose of embedding hidden clues within the text to suggest its provenance. daringfireball.net/2026/08/anthro…

I mean this in the kindest possible way; this is an almost unimaginably bad take.

Because apart from botching all the technical details, your objections apply to *all* LLM "writing", not to watermarked writing in particular.

But my objection does not apply to all LLM writing. LLM output can be better or worse. I am more convinced today than a week ago that SynthID watermarking makes it worse. Maybe only slightly worse. But it has to be worse. That’s how it works.

First, thanks, that's genuinely appreciated. Neither flu or potty-training is particularly bad as these things go, but both are discombobulating. As is the threat of imminent redundancy, which I keep forgetting about. Okay, I think the problem is that - even granting for argument's sake that... 1/

Meant to thread this here; bsky.app/profile/elto...

"Botching all the technical details" is not really fair, looking at it again. Apologies. We're potty-training at the moment, and I have flu, and judging the energy and emotion of the moment comes less easily than usual.