@revolutia.ai — 8x breakout

Loading video…

Anthropic acaba

@revolutia.ai

151.4K followers · usually 40.3K views a post

Views

322.7K

Likes

24.8K

Comments

129

2.7K

Promoted product

Watermarks-remover

Why it worked

The video taps into the growing concern and discussion around AI-generated content and the desire to bypass detection. By presenting a practical solution to a relevant problem, it offers immediate value and encourages engagement.

Summary

The video explains how Anthropic's AI model, Claude, embeds hidden signatures in its generated files using invisible Unicode and metadata. A tool called 'Watermarks-remover' was developed to remove these signatures from various file types, including those generated by ChatGPT and Gemini, allowing users to bypass AI detection.

Structure

1Introduction to Claude's hidden signatures
2Development of Watermarks-remover tool
3Demonstration of tool's capabilities
4Cross-compatibility with other AI models
5Call to action to access the tool

Product placement

Anthropic: ACTED - it is the company that generated the file, and its name appears multiple times. Removing it would change what happens in the video. Claude: ACTED - it is the AI model that generated the file, and its name appears multiple times. Removing it would change what happens in the video. ChatGPT: ACTED - it is an AI model that generated a file, and its name appears. Removing it would change what happens in the video. Gemini: ACTED - it is an AI model that generated a file, and its name appears. Removing it would change what happens in the video. Watermarks-remover: ACTED - it is a tool that removes watermarks from files. Removing it would change what happens in the video.

Original caption

Cada archivo que genera Claude lleva una firma oculta dentro. Y la mayoría de la gente no lo sabe. No es un logo en la esquina. Es unicode invisible y metadatos firmados dentro de tus PNG, PDF, SVG, HTML. Todo lo que sale del chat, de Cowork o de Claude Code. 24 horas después ya había repos publicados para borrarla. Y no solo de Claude: también de ChatGPT y Gemini. Lo que más me llama la atención no es la tecnología. Es lo rápido que se organizó la gente para esconder algo que, en teoría, no tenía nada de malo. Porque ahí está la pregunta de verdad: ¿por qué te incomoda que se sepa? Si el trabajo es bueno, la firma no te quita nada. Si el trabajo es malo, borrar los metadatos no lo arregla. Y a nivel práctico: esto va a empezar a aparecer en contratos. Cláusulas de divulgación, entregables con procedencia verificable, clientes que preguntan. El que hoy esconde va a tener un problema el día que se lo pidan por escrito. Mi apuesta es la contraria. Di que usas IA, cobra por el criterio, y deja de competir con los que venden horas.