Some Claude users are mad that Anthropic’s new watermarks will catch them using it at their jobs, classes

This won’t work - you can always use a dumber, open model to rephrase and remove the watermark. All this does is marketing for anthropic, which I see most people are buying. Gives you a false impression you will easily be able to figure out if it’s done by AI, which it is not. Liars are still gonna lie, cheaters still gonna cheat.

Sure, but they need the watermarking to later claim the copyright on the stuff that people do create in the open with their agents.

“They would never do that!”

“And even if they did, I would sue them, and then the justice system would…”

“uh… uh oh”

My tinfoil head tells me they don’t want to ingest their own slop for training. But this only helps them if others do the same. Data poisoning is a real problem for llm future.