Tools

Search

Explore

Videos Channels Figures

Atmrix

About

Tools

Search

Explore

Videos Channels Figures

Atmrix

About

EP

Ethan Perez

02/28/25

@ Anthropic

The constitution enumerates categories of requests deemed harmful, such as making weapons of mass destruction, and harmless, like writing poems.

Video

A

Defending against AI jailbreaks

@ Anthropic

02/28/25

Related Takeaways

EP

Ethan Perez

02/28/25

@ Anthropic

Constitutional classifiers use a natural language set of rules to specify harmful and harmless topics, guiding the model's responses accordingly.

TB

Trenton Bricken

05/23/25

@ Dwarkesh Patel

It's better for the Constitution to specify actions, like protecting free speech, rather than just aiming for human flourishing, which is too vague.

MA

Marc Andreessen

01/26/25

@ Lex Fridman

The US First Amendment has multiple carve-outs, meaning that certain types of speech, like incitement to violence and child pornography, are not protected.

EP

Ethan Perez

02/28/25

@ Anthropic

The flexibility of our approach allows us to quickly adapt to new threats by simply rewriting the constitution to include new harmful categories.

MS

Mustafa Suleyman

09/04/23

@ The Diary Of A CEO

We must restrict access to dangerous technologies to prevent misuse, even if it frustrates those who want to innovate.

A

Anthropic Cast

02/28/25

@ Anthropic

The constitutional classifiers approach is extremely flexible and can be easily updated to respond to various novel threats.

SH

Steve Huffman

09/26/24

@ Sequoia Capital

The first content policy was like a convention center where various communities could gather, but we had to draw a line against harmful content.

MA

Marc Andreessen

01/26/25

@ Lex Fridman

The original discussions around censorship focused on hate speech and misinformation, with hate speech being the first topic addressed.

PH

Peter H. Diamandis Cast

06/04/25

@ Peter H. Diamandis

Intelligent weapons should differentiate between targets to minimize civilian casualties, but this raises concerns about the assumption of benevolence in their use.