Alex Cui

@alexcdot

Uncovering the world's slop at @GPTZeroAI | Staying curious

Toronto, Ontario

alexcui.ca

Replies Reposts Media

Pinned
Alex Cui
@alexcdot
Aug 11
Claude's watermark probably doesn't work how you think. As the CTO of GPTZero, I'll explain how Anthropic, Google and OpenAI are building text watermarking in this brief explainer and whether it can be defeated. Almost all forms of watermarking that are fast and cheap enough for
NIK
@ns123abc
Aug 10
🚨 JUST IN: Claude models will now have invisible watermarks embedded in ALL text, and ALL metadata attached to files…
290
Alex Cui
@alexcdot
Aug 12
While this is a hilarious take, there's something deeper going on I think the reason why people don't like claude watermark is because it makes it seem like claude gets all the credit for the text when, in many cases, people supplied the ideas/voice/draft to get the final output
GIF
viviolet
@violetrose1410
Aug 11
"I don't want invisible watermarks embedded in my content" then........ write it?? Yourself????
13
Alex Cui
@alexcdot
Aug 11
Can we take a moment to appreciate that they dropped this... from their support page? Anthropic totally rewriting the B2B SaaS playbook once again
NIK
@ns123abc
Aug 10
🚨 JUST IN: Claude models will now have invisible watermarks embedded in ALL text, and ALL metadata attached to files…
3
Alex Cui
@alexcdot
Aug 11
Claude: We are watermarking your tokens Me: No problem, I'll use Kimi! Kimi: We distilled from Claude so it's still watermarked Me: No problem, I'm switching to my LLM trained only on pre-1931 text
14
Alex Cui
@alexcdot
Jul 29
Our team just revealed how @PwC pushed multiple reports with awful hallucinations as "thought-leadership" and got away with it for years. The scary thing is your AI agents are citing their slop as truth, putting you at risk. For instance, PwC claimed @jpmorgan used Agentic AI
4