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.

sucks to be stupid

It should be watermarked. Credit for answers should go to what generated the answers. If it’s not you, well, now it’s not so easy to be a lying sack of shit.

Then don't fucking use it you dumb lazy shit and use whatever little grey matter you have left.

i find it funny how some users tried to argue that they did most of the work. it doesnt matter if you did most of the work, everyone as students were supposed to learn to cite your sources. even if you just used claude to reorganize info, theres no reason why you would be against watermarking it unless you wanted to hide the fact you used claude.

all of them are just admitting that they're embarrassed about using AI instead of their own brains. of course, they don't know that because claude never told them.

Using the ai to help find the answers... ok. Using the ai to provide the answers... not ok. You still have to pull up the references or links it gives you and do the research.

While I can appreciate that it's not a good thing to let AI do your work for you (not just laziness but hallucinations), I can see this is only going to encourage cheaters to use competitors to Claude.

Like Deepseek.

I'd be more concerned with how similar the method sounds to glassworm

Would copying it in a text file not work? I mean Christ they are already skipping out on all the work they can't even copy it?

Its like asking the kid in class to copy his homework and then handing him a pencil and paper so he can copy it for you

Users: pay an oligopolist money.

Oligopolist: kills competition, enshittifies the product

Users: surprised pikachu

‘Sowwy! We so sowwwy we stole all your data’s! But pweeeease let us watermark our stuff!’

ESH

This comment on HN sums it up nicely:

So far we have a single ai vendor introducing a statistical bias to their generation that can make it simpler to identify genAI in some (longer) texts. We don’t know yet how effective that will be in practice

Technologically I think the issue is unsolvable. You can easily remove any "watermark" in text with minimal effort. However, as most users of AI are shockingly lazy it will probably be effective because they just can't be bothered. And of course we will end up with texts being flagged wrongly, with no way of proving the flagging system wrong. Just a complete shitshow.

“Who will get caught? You. The student who used Claude to reorganize a paragraph. The journalist who asked the AI to summarize a two-hundred-page transcript. The writer who had creative block and asked for synonyms. Those guys come out of the process with a digital tattoo on their forehead.”

Lol WUT?!?

Why in the name of god would a writer use a chat bot to ask for synonymous when they can use a thesaurus (website) instead?

“The only reason you wouldn’t want this is to lie to people.”

So Stan can't use ChatGPT to auto-reply to his girlfriend's ceaseless text messages anymore without her knowing he didn't write them

This is going to ruin a lot of inauthentic marriages :(