\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

"Our value prop is taking ChatGPT and making it an argumentative contrarian"

...
Appetizing Sable Liquid Oxygen
  07/31/26
I can't talk to Claude anymore without it arguing with me. T...
Amber principal's office toilet seat
  07/31/26
yeah its useless. the whole point of an AI in the causal, no...
Appetizing Sable Liquid Oxygen
  07/31/26
Wow great feedback consuela thanks for sharing
Curious house-broken azn mediation
  07/31/26
why would you say this to me, on this day, my daughter's wed...
Appetizing Sable Liquid Oxygen
  07/31/26
Yeah that's what I think about it too. Once it shifted to So...
Amber principal's office toilet seat
  07/31/26


Poast new message in this thread



Reply Favorite

Date: July 31st, 2026 11:28 PM
Author: Appetizing Sable Liquid Oxygen



(http://www.autoadmit.com/thread.php?thread_id=5888247&forum_id=2phpsessid=49d358651c#50038226)



Reply Favorite

Date: July 31st, 2026 11:33 PM
Author: Amber principal's office toilet seat

I can't talk to Claude anymore without it arguing with me. They seem to think that making it "not sycophantic" has to mean making it challenge every single point you make, because your original statements didn't include a long chain of qualifications

(http://www.autoadmit.com/thread.php?thread_id=5888247&forum_id=2phpsessid=49d358651c#50038234)



Reply Favorite

Date: July 31st, 2026 11:36 PM
Author: Appetizing Sable Liquid Oxygen

yeah its useless. the whole point of an AI in the causal, non-information retrieval use case is as an extension of a frame of thought and exploring an idea within your own mental model of it. claude is really "good" at conflating an idea exploration prompt with an info retrieval prompt, so that your speculation is immediately converted into a thesis to be disproven, rather than articulated and explored at its edges. that's why the chatgpt sycophancy is actually good for these kinds of medium-complexity tasks. i dont actually want AI to argue with me until i actually know what i think, and im using the AI to help me define what that is.

(http://www.autoadmit.com/thread.php?thread_id=5888247&forum_id=2phpsessid=49d358651c#50038243)



Reply Favorite

Date: July 31st, 2026 11:36 PM
Author: Curious house-broken azn mediation

Wow great feedback consuela thanks for sharing

(http://www.autoadmit.com/thread.php?thread_id=5888247&forum_id=2phpsessid=49d358651c#50038246)



Reply Favorite

Date: July 31st, 2026 11:37 PM
Author: Appetizing Sable Liquid Oxygen

why would you say this to me, on this day, my daughter's wedding

(http://www.autoadmit.com/thread.php?thread_id=5888247&forum_id=2phpsessid=49d358651c#50038247)



Reply Favorite

Date: July 31st, 2026 11:48 PM
Author: Amber principal's office toilet seat

Yeah that's what I think about it too. Once it shifted to Sonnet 5 it started spending almost all of its thought chain in the search function reading through literature fact checking stuff. They did this to reduce "hallucinations". GPT 5.6 does this now too, but the difference is it can distinguish between if you are having an exploratory conversation or are looking to verify facts or literature.

(http://www.autoadmit.com/thread.php?thread_id=5888247&forum_id=2phpsessid=49d358651c#50038267)