@adamstewartmarketing: Anthropic just announced that every Claude model made after August 2nd hides a secret mark inside the text it gives you. #anthropic #claude #ai
Don’t tell us what we already know ! WHAT IS THE SOLUTION ???? That’s what matters
2026-08-17 01:36:27
0
alaskabiker907 :
Anthropic is going to start losing a lot of money over this. Major shift in their algos.
2026-08-14 07:36:42
15
HammOnRye :
I don’t understand, if AI writing is modeled after human writing, why wouldn’t my human writing also carry the watermark? There aren’t magic words that only Claude would use…
2026-08-14 23:40:27
4
George :
ok... read it out to a Google doc on voice transcription
2026-08-12 17:25:53
15
Joey :
copy paste notepad = no watermark
2026-08-13 03:10:47
1
malcolmx996 :
that's not how it works
2026-08-12 21:47:22
5
Zamozi :
I mean I have to read over it anyway so I’ll just voice to text it into another doc.
2026-08-12 22:05:03
1
Hitchpin :
lol watermark that no one knows then how can checkers know? I assume it’s common words but couldn’t you just dolly chain it through other AIs?
2026-08-12 23:24:10
3
Steve :
As soon as they give a way to read it there will be 10,000 apps to remove it.
2026-08-13 05:00:22
3
MFT_V_ :
Windows notepad POV: It's my time to shine
2026-08-12 19:47:18
14
imrohanzaki :
screenshot get ocr again save as BMP then open in paint and save as jpg or png mark cleared
2026-08-12 22:37:48
6
AhhhFiddleSticks :
Or you could just have an open weights model paraphrase everything for you
2026-08-13 00:42:45
1
unknown :
Take what it gives you and rewrite it in your own words, word by word? Thought that’s what everyone did anyways 😅
2026-08-13 21:38:44
1
Kevin Prince843 :
Give it a year and everything will have one word of AI making the mark efficiently useless.
2026-08-13 01:45:30
2
vdonevska :
Good generative AI didn't invent clear structure, transitions, sophisticated vocabulary, balanced arguments, or concise explanations. Humans did. AI learned those patterns from human writing. So when an AI detector treats polished prose as evidence of AI authorship, it can end up detecting qualities of competent human writing rather than reliably detecting who produced it.
2026-08-12 22:01:48
4
GoodMorning :
So now literally we need to write 200 cover letters to apply for a job because ATS will check if the letter is AI written. As if being jobless isn't bad enough, applying for jobs just got worse.
2026-08-13 01:07:34
1
Maximus :
OCR tools are the solution
2026-08-12 21:57:42
3
canned can :
Do that enough… make it harder to use and that shet and the east will win the ai race. The west is so stupid building constraints, regulations and this shit while china is flying … dont be surprised when we have to learn chinese
2026-08-12 20:05:55
1
zer0conf :
just use other ai providers
2026-08-12 23:27:02
1
AI & Automations :
we had this with chatgpt a while back 😂😂😂
2026-08-12 16:41:43
1
Avolang_Beratung :
Aaaaand they will implement this into their older models at any moment as well :)
2026-08-12 20:53:08
1
Qwerty :
Then another company will built an AI to correct it!! It’s too late
2026-08-12 19:53:40
1
mokokolord :
just type it manually
2026-08-13 09:42:56
0
محمد عـبـدالـعـزيـز :
Its a lie. Putting text no one can read…. Who believes this?
2026-08-13 12:43:39
0
AI New Wave :
There is another free tool to detect Claude watermarks but easier its called freeaiden.
2026-08-15 07:33:34
0
To see more videos from user @adamstewartmarketing, please go to the Tikwm
homepage.