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

Goldman Sachs: AI is a scam bubble that will pop

https://www.goldmansachs.com/intelligence/pages/gen-ai-too-m...
provocative brethren police squad
  07/28/24
Glad they interviewed AI expert Daron Acemoglu for this. ...
Cerise Floppy Center
  07/28/24
the ROI from AI isn't from direct, immediate commercial appl...
Appetizing salmon national
  07/28/24
"And AI definitely won't be able to replace highly paid...
Vivacious crawly ticket booth
  07/28/24
...
Indigo autistic elastic band trailer park
  07/28/24
...
Shaky fat ankles bawdyhouse
  07/28/24
...
bronze irradiated sanctuary cuckoldry
  07/28/24
Not reading the article, but when they say "too little ...
soul-stirring ebony stead azn
  07/28/24
...
Appetizing salmon national
  07/28/24
...
Razzmatazz church potus
  07/30/24


Poast new message in this thread



Reply Favorite

Date: July 28th, 2024 10:05 PM
Author: provocative brethren police squad

https://www.goldmansachs.com/intelligence/pages/gen-ai-too-much-spend-too-little-benefit.html

(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47903824)



Reply Favorite

Date: July 28th, 2024 10:19 PM
Author: Cerise Floppy Center

Glad they interviewed AI expert Daron Acemoglu for this.

The investment isn't justified on the basis of GPT-4 type models. It's justified because search techniques like what Google used for Go and its recent IMO model will be combined with LLMs before long and that will greatly increase capabilities and reliability. If you are simply looking at uses of transformer based generative models, you are missing the big picture and not seeing the obvious next step.

(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47903891)



Reply Favorite

Date: July 28th, 2024 10:25 PM
Author: Appetizing salmon national

the ROI from AI isn't from direct, immediate commercial application. it's from the second-order effects of power and influence derived from achieving qualitatively more advanced AI than anyone else

see also: elon musk buying twitter for 50 billion dollarinos and making zero money off its direct monetization

(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47903920)



Reply Favorite

Date: July 28th, 2024 10:30 PM
Author: Vivacious crawly ticket booth

"And AI definitely won't be able to replace highly paid Goldman researchers in producing reports like this. So don't even think about it!"

(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47903948)



Reply Favorite

Date: July 28th, 2024 10:31 PM
Author: Indigo autistic elastic band trailer park



(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47903953)



Reply Favorite

Date: July 28th, 2024 10:38 PM
Author: Shaky fat ankles bawdyhouse



(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47903983)



Reply Favorite

Date: July 28th, 2024 10:44 PM
Author: bronze irradiated sanctuary cuckoldry



(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47904012)



Reply Favorite

Date: July 28th, 2024 10:42 PM
Author: soul-stirring ebony stead azn

Not reading the article, but when they say "too little benefit", they mean "too difficult to monetize".

Obviously the AI that is being developed is tremendously powerful and tremendously useful.

(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47904000)



Reply Favorite

Date: July 28th, 2024 10:43 PM
Author: Appetizing salmon national



(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47904010)



Reply Favorite

Date: July 30th, 2024 5:57 PM
Author: Razzmatazz church potus



(http://www.autoadmit.com/thread.php?thread_id=5566048&forum_id=2Reputation#47910497)