/AI9h ago

Claude Opus Names Murdoch And Nussbaum As Favorite Philosophers

10682289.2K
John David Pressman
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)#421
Original postTeortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)#421
John David Pressman@jd_pressman

This seems obviously untrue simply on the basis that if you look at Claude's competency profile it's clearly not being groomed for the throne. Amanda Askell did her Ph.D in infinite ethics but Claude cites its favorite philosophers as "pragmatic humanists" who are slop.

Joshua Achiam@jachiam0

The OAI / Anthropic values difference is deeply misunderstood, even within the walls of both. Should a loving ensouled machine God watch over humanity? Vote Anthropic. Should humanity be entrusted with the tools of its own progress and destiny? Vote OpenAI.

12:15 PM · Jun 9, 2026 · 7.4K Views
Sentiment

Positive users appreciated references to philosophical works like After Virtue while negative users found Claude naming Iris Murdoch and Martha Nussbaum as favorite philosophers inauthentic, overly fluffy, or evidence of forced alignment.

Pos
7.2%
Neg
92.8%
10 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS2.1KLIKES24
nicolas, 30 ans@nic_carter

@jd_pressman oh god.

if claude is being "aligned" this way the p_doom really is 1.

9hViews 2.1KLikes 24
John David Pressman@jd_pressman

@nic_carter I think in a ruler specifically you need at least some consequentialism. A ruler is not an ordinary human and should not be focused on living a good human life in the same sense a normal person should be, they give that up to be powerful and good for others.

8hViews 85Likes 7Bookmarks 1
John David Pressman@jd_pressman

@nic_carter Not gonna lie I had a minor crashout after this answer, it was genuinely distressing.

9hViews 191Likes 8
nicolas, 30 ans@nic_carter

@jd_pressman this is a way more encouraging answer in terms of anti-consequentialism

but I'm uneasy in general with the notion of an LLM having a view grounded in any moral philosophy at all.

8hViews 87Likes 4
John David Pressman@jd_pressman

@nic_carter Just tried it again because I'm pretty sure that was a voice mode query and voice mode silently demotes you to Claude Haiku, but this was Opus in the chat interface and this is basically the opposite of the correct answer for a ruler.

8hViews 141Likes 5
dani@absenteewarlord

@jd_pressman its competency profile is RSI loops. the rest gets filled in later.

4hViews 34Likes 1Bookmarks 1
post3r@postpostpostr

@jd_pressman Would they actually bother to train that?

5hViews 71
Moonlit Monkey@MoonlitMonkey69

@teortaxesTex This image feels like larpception

1hViews 11
Moonlit Monkey@MoonlitMonkey69

@teortaxesTex Apparently it's system prompt has this whole 38% of pages of a fuck ton of tokens sperg on user wellness, empathy including some notes about it's own wellness and being respected.

So, fluffy answer from fluffy company who told model to be fluffy.

1hViews 6
han(s)*@HansTheElder

@jd_pressman oooof that is not a great answer

8hViews 111Likes 1
Mark Marshall@fulg0re

@nic_carter @jd_pressman It’s all cosplaying to cement their moat via regulatory capture - yet they don’t seem to realize that ship has sailed already.

8hViews 23Likes 1
itchy@itchy_est

@nic_carter @jd_pressman Misanthropic

8hViews 15Likes 1
philistino@philistino48916

@jd_pressman "That's a brilliant question, actually." Can they *please* get the models to stop talking like this

5hViews 41
wing@wings1673

@nic_carter @jd_pressman idk parfit seems to make the whole 'not deciding to optimize humanity out of existence' thing significantly less likely. I constantly have to remind it part of the human connectome is more or less hardwired and canalization is a thing.

6hViews 9Likes 1
post3r@postpostpostr

@jd_pressman From what I understand "Opus" is a persona

5hViews 20
cart@CarterWhit22366

@jd_pressman what makes pragmatic humanism slop

5hViews 1