The most prestigious law school admissions discussion board in the world.
|
 |
 
|
6/16/25 AI thread (huge dump of updates)
Poast new message in this thread
Date: June 16th, 2025 10:22 AM Author: Total Jewish Pumo Death Trump administration's plans to integrate AI throughout the federal government leaked:
https://x.com/kimmonismus/status/1934359981031366823
Nvidia reveals roadmap for physical AI and massive AI factories:
https://x.com/heyshrutimishra/status/1933216576721060332
Models are now being made that can create their own synthesized data and update their own weights:
https://x.com/jyo_pari/status/1933350025284702697
Theory of mind primer on how LLMs work and why:
https://nostalgebraist.tumblr.com/post/785766737747574784/the-void
LLMs using sub agents are much more effective than a single LLM agent acting alone. There is also a paper on Anthropic's website about this:
https://x.com/adonis_singh/status/1933848838189531448
Good article on why the human brain is much more effective if it has actually internalized information, instead of having to look it up (strong argument against society embracing LLMs):
https://x.com/gtdad/status/1933939935612182886
Paper on how human brains atrophy when they use AI heavily on a regular basis:
https://x.com/MoonL88537/status/1934559050806485229 |
Date: June 16th, 2025 10:23 AM Author: Oh, you travel? ( ) ty TJPD (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49019934) |
 |
Date: June 16th, 2025 10:56 AM Author: Total Jewish Pumo Death (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49020053) |
|
Date: June 16th, 2025 10:58 AM Author: Police Boner (🧐) (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49020064) |
|
Date: June 16th, 2025 4:10 PM Author: internet g0y (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49021337) |
|
Date: June 16th, 2025 10:30 AM I trained a local LLM on xo and it called me a fag. (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49019950) |
|
Date: June 16th, 2025 10:57 AM Author: Police Boner (🧐) Models are now being made that can create their own synthesized data and update their own weights:
https://x.com/jyo_pari/status/1933350025284702697
holy shit here we goooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
|
|
Date: June 16th, 2025 11:00 AM Author: Total Jewish Pumo Death This seems like a pretty big deal imo there are several groups working on this right now (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49020070) |
|
Date: June 16th, 2025 11:03 AM Author: Total Jewish Pumo Death Explanation for why AI is only allowed to code for 7 hours a day due to labor laws:
https://x.com/ben_j_todd/status/1934284189928501482 |
|
Date: June 16th, 2025 4:11 PM Author: scholarship (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49021344) |
|
Date: June 16th, 2025 11:07 AM Author: Total Jewish Pumo Death OpenAI guy implicitly admitting that LLM progress has stalled and that the focus is now on creating imaginary friends/girlfriends/boyfriends as a commercial application to make money from users:
https://x.com/vitrupo/status/1934627428372283548 |
|
Date: June 16th, 2025 11:11 AM Author: Police Boner (🧐) This guy's resume is like a cowgod wet dream. Gave up on computer science and became a Liberal Artist Product Manager of the most Important Project of Our Time
Google DeepMind Google DeepMind Full-time · 6 mosFull-time · 6 mos Group Product Manager Group Product Manager May 2025 - Present · 2 mosMay 2025 to Present · 2 mos Building the world’s best AI developer platform.Building the world’s best AI developer platform. Senior Product Manager Senior Product Manager Jan 2025 - May 2025 · 5 mosJan 2025 to May 2025 · 5 mos Helped us move over to Deepmind, still lead product for Google AI Studio. Helping build the Gemini API into the best platform in the world for developers to build with AI.
Harvard University Master of Liberal Arts, Extension Studies, Digital Media DesignMaster of Liberal Arts, Extension Studies, Digital Media Design Grade: 3.91/4.0Grade: 3.91/4.0
Focus on learning design, software engineering, and technology. Harvard University Bachelor of Liberal Arts, Extension Studies, Cum Laude, Computer ScienceBachelor of Liberal Arts, Extension Studies, Cum Laude, Computer Science Grade: 3.53/4.00Grade: 3.53/4.00 Area of study: Computer Science
University of Oxford Undergraduate Advanced Diploma in IT System Analysis and Design, Information TechnologyUndergraduate Advanced Diploma in IT System Analysis and Design, Information Technology 2019 - 20212019 - 2021 Grade: DistinctionGrade: Distinction
De Anza College Computer Science to transferComputer Science to transfer Grade: 3.49/4.00Grade: 3.49/4.00 Activities and societies: Phi Theta Kappa Honor SocietyActivities and societies: Phi Theta Kappa Honor Society
Buffalo Grove High School High School DiplomaHigh School Diploma Grade: 4.79/5.00 |
|
Date: June 16th, 2025 11:15 AM Author: Total Jewish Pumo Death "Harvard University
Master of Liberal Arts, Extension Studies, Digital Media DesignMaster of Liberal Arts, Extension Studies, Digital Media Design"
Lmao
DIGITAL MEDIA. DESIGNMASTER. OF. LIBERAL. ARTS. |
|
Date: June 16th, 2025 12:06 PM Author: Total Jewish Pumo Death Political actors are already realizing the power of AI to re-write history:
https://x.com/braxton_mccoy/status/1934593157632577985 |
|
Date: June 16th, 2025 4:09 PM Author: Total Jewish Pumo Death this essay in particular (also in OP) was really good. a must read for people interested in the theory of mind of LLMs
https://nostalgebraist.tumblr.com/post/785766737747574784/the-void
couple of other related links:
https://www.lesswrong.com/posts/gTZ2SxesbHckJ3CkF/transformers-represent-belief-state-geometry-in-their (not sure if i agree with their conclusions, need to think about it more)
https://www.lesswrong.com/posts/vJFdjigzmcXMhNTsx/simulators
this is a highly recommended post but i haven't read it myself yet |
|
Date: June 16th, 2025 4:10 PM Author: scholarship Interesting ty (http://www.autoadmit.com/thread.php?thread_id=5738495&forum_id=2#49021341) |