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

Starting to come around to the "AI is overhyped garbage" side

...
drunken azure shrine foreskin
  06/10/25
wow what caused this
fragrant brindle prole
  06/10/25
AI told me this
drunken azure shrine foreskin
  06/10/25
It's a very useful tool for some things (coding, learning ab...
mewling multi-billionaire
  06/10/25
The limitations and errors and overconfidence (this is a pet...
fantasy-prone buck-toothed stage milk
  06/10/25
Most of the complaints about ChatGPT apply about 10x more to...
Multi-colored olive step-uncle's house depressive
  06/10/25


Poast new message in this thread



Reply Favorite

Date: June 10th, 2025 5:45 PM
Author: drunken azure shrine foreskin



(http://www.autoadmit.com/thread.php?thread_id=5735793&forum_id=2...id.#49003139)



Reply Favorite

Date: June 10th, 2025 5:46 PM
Author: fragrant brindle prole

wow what caused this

(http://www.autoadmit.com/thread.php?thread_id=5735793&forum_id=2...id.#49003142)



Reply Favorite

Date: June 10th, 2025 5:51 PM
Author: drunken azure shrine foreskin

AI told me this

(http://www.autoadmit.com/thread.php?thread_id=5735793&forum_id=2...id.#49003152)



Reply Favorite

Date: June 10th, 2025 6:00 PM
Author: mewling multi-billionaire

It's a very useful tool for some things (coding, learning about a topic in an introductory way, writing rough drafts/outlines of papers, revising existing papers), but has obvious and clear limitations when it comes to actually reasoning and generalizing from principles.

(http://www.autoadmit.com/thread.php?thread_id=5735793&forum_id=2...id.#49003174)



Reply Favorite

Date: June 10th, 2025 6:05 PM
Author: fantasy-prone buck-toothed stage milk

The limitations and errors and overconfidence (this is a pet peeve of mine with humans too; has your ai ever been like “sorry, I don’t know”? no, it just feeds you the wrong answer with the same delivery as ones where it has the facts down cold) have been grating on me personally lately.

(http://www.autoadmit.com/thread.php?thread_id=5735793&forum_id=2...id.#49003182)



Reply Favorite

Date: June 10th, 2025 7:36 PM
Author: Multi-colored olive step-uncle's house depressive

Most of the complaints about ChatGPT apply about 10x more to humans. It’s only real limitation is not having enough context to understand the real ask. It either doesn’t know about relationship dynamics or what length or style is preferred by the reader. But it is far faster than everyone and more accurate and compelling than most. The main thing that is annoying about it is that it is overly servile and so idiots that command it can do real damage because it will turn a niche or minority viewpoint into something that sounds mainstream or market. It also reinforces these idiots opinions so they get even more emboldened by their wrong viewpoint

(http://www.autoadmit.com/thread.php?thread_id=5735793&forum_id=2...id.#49003400)