LessWrong (30+ Karma)

“Open Problems With Claude’s Constitution” by Zvi


Listen Later

The first post in this series looked at the structure of Claude's Constitution.

The second post in this series looked at its ethical framework.

This final post deals with conflicts and open problems, starting with the first question one asks about any constitution. How and when will it be amended?

There are also several specific questions. How do you address claims of authority, jailbreaks and prompt injections? What about special cases like suicide risk? How do you take Anthropic's interests into account in an integrated and virtuous way? What about our jobs?

Not everyone loved the Constitution. There are twin central objections, that it either:

  1. Is absurd and isn’t necessary, you people are crazy, OR
  2. That it doesn’t go far enough and how dare you, sir. Given everything here, how does Anthropic justify its actions overall?
  3. The most important question is whether it will work, and only sometimes do you get to respond, ‘compared to what alternative?’

    Post image, as chosen and imagined by Claude Opus 4.5

    Amending The Constitution

    The power of the United States Constitution lies in our respect for it, our willingness to put it [...]

    ---

    Outline:

    (01:30) Amending The Constitution

    (03:45) Details Matter

    (05:09) WASTED?

    (07:40) Narrow Versus Broad

    (09:00) Suicide Risk As A Special Case

    (10:36) Careful, Icarus

    (11:19) Beware Unreliable Sources and Prompt Injections

    (12:15) Think Step By Step

    (12:50) This Must Be Some Strange Use Of The Word Safe I Wasn't Previously Aware Of

    (16:26) They Took Our Jobs

    (20:08) One Man Cannot Serve Two Masters

    (24:29) Claude's Nature

    (30:14) Look What You Made Me Do

    (32:32) Open Problems

    (36:40) Three Reactions and Twin Objections

    (36:57) Those Saying This Is Unnecessary

    (38:05) Those Saying This Is Insufficient

    (39:56) Those Saying This Is Unsustainable

    (43:12) We Continue

    ---

    First published:

    January 28th, 2026

    Source:

    https://www.lesswrong.com/posts/vFAJxua3Qc6S8MbqG/open-problems-with-claude-s-constitution

    ---

    Narrated by TYPE III AUDIO.

    ---

    Images from the article:

    Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

    ...more
    View all episodesView all episodes
    Download on the App Store

    LessWrong (30+ Karma)By LessWrong


    More shows like LessWrong (30+ Karma)

    View all
    The Daily by The New York Times

    The Daily

    113,122 Listeners

    Astral Codex Ten Podcast by Jeremiah

    Astral Codex Ten Podcast

    132 Listeners

    Interesting Times with Ross Douthat by New York Times Opinion

    Interesting Times with Ross Douthat

    7,266 Listeners

    Dwarkesh Podcast by Dwarkesh Patel

    Dwarkesh Podcast

    529 Listeners

    The Ezra Klein Show by New York Times Opinion

    The Ezra Klein Show

    16,315 Listeners

    AI Article Readings by Readings of great articles in AI voices

    AI Article Readings

    4 Listeners

    Doom Debates by Liron Shapira

    Doom Debates

    14 Listeners

    LessWrong posts by zvi by zvi

    LessWrong posts by zvi

    2 Listeners