The most prestigious law school admissions discussion board in the world.
|
 |
 
|
the ai bubble is popping
Poast new message in this thread
Date: August 3rd, 2026 1:57 AM Author: screen memories (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042021) |
Date: August 3rd, 2026 2:13 AM Author: peaks but not a pop tp (Prestige Feuds) Wait (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042027) |
|
Date: August 3rd, 2026 2:14 AM Author: The Penis peaking not popping (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042028) |
 |
Date: August 3rd, 2026 2:14 AM Author: peaks but not a pop tp (Prestige Feuds) !!! (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042029) |
|
Date: August 3rd, 2026 2:54 AM Author: animeboi (.) it's over. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042039) |
|
Date: August 3rd, 2026 3:42 AM Author: .,.,...,..,.,.,:,,:,...,:::,...,:,.,.:..:. OpenAI just casually announced proofs on Saturday for conjectures each of which would likely win someone a Fields medal. there is no bubble here. it's the end of human intellectual supreamcy. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042046) |
|
Date: August 3rd, 2026 7:23 AM Author: have you ever looked at a quarter Oh, the gdp implications! (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042161) |
|
Date: August 3rd, 2026 9:12 AM Author: The Penis I think the deprecation of the human mind is a little bigger than current US GPT implications. Just sayin. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042270) |
|
Date: August 3rd, 2026 6:13 PM Author: .,.,...,..,.,.,:,,:,...,:::,...,:,.,.:..:. recursive self-improving LLMs will solve continual learning and then automate every other job. the current generation of models is capable of doing AI research experiments at scale. likely <3 years from AGI, hth. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50043193) |
|
Date: August 3rd, 2026 9:48 AM Author: UhOh that's cool but how are they going to make money? the chinks will have one next month that does the same thing for free. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042339) |
|
Date: August 3rd, 2026 10:07 AM Author: The Penis Still haven't used any Chinese AI that wasn't pure trash (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042377) |
|
Date: August 3rd, 2026 6:25 PM Author: the walter white of this generation (walt jr.) Lol jfc. "Proofs" that "would likely" win someone a "field medal"?! Wow! (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50043207) |
|
Date: August 3rd, 2026 6:29 PM Author: Jared Baumeister (no Fields medals were awarded) (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50043214) |
|
Date: August 3rd, 2026 6:36 PM Author: LathamTouchedMe What about this argument that while impressive, it's applicability is a lot narrower than the AI prophets would have you believe.
"There's a specific subset of mathematical problems which can be disproven via counterexamples, which take humans a long time to map and calculate. One of LLMs unique capabilities is that it can produce small, relatively simple programs at-scale, and because these problems are so well articulated and their potential solutions already well understood they lend themselves to this capability." |
|
Date: August 3rd, 2026 7:00 PM Author: .,.,...,..,.,.,:,,:,...,:::,...,:,.,.:..:. These aren't all counterexamples. 7 out of 10 weren't.
It's also worth noting there has been a tendency with math AI to make arguments of the sort, "well it's good at this <problem class>, but it's because this is well suited to AI", and then see the next generation of math AIs tackle a much broader range of problems that were supposedly uniquely human. This is essentially a deluded cope that relies on someone to ignore the rapid improvement rate that is largely driven by more compute and training. It's guaranteed at this point the models in a few months will solve many more problems than they can now. There is nothing special about problems outside of the current capability range. They just require more training or inference compute to solve. |
|
Date: August 3rd, 2026 9:21 AM Author: law is ea$y (🧐) pooping* (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042283) |
|
Date: August 3rd, 2026 9:41 AM Author: .,.,.,.,.,,,.,.,.,.,., not even close, my op. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042329) |
|
Date: August 3rd, 2026 9:47 AM Author: chandler (retired) I used Sol 5.6 yesterday and it was such a major step up from any previous open AI versions I used I think we’re just getting started. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042334) |
|
Date: August 3rd, 2026 9:49 AM Author: The Penis Codex sol ultra is so wildly good it should be considered AGI. If you had enough compute you could basically do anything with it with little human direction. Some mathematics superstar just joined open ai after playing with it, because of how good it is at testing and eliminating conjectures. (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042342) |
|
Date: August 3rd, 2026 9:58 AM Author: law is ea$y (🧐) mistook this as chilmata for a second and got very excited (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50042361) |
|
Date: August 3rd, 2026 6:38 PM Author: Shema Yisrael lmao it's been out for weeks yngmi (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50043236) |
|
Date: August 3rd, 2026 9:49 AM Author: The Dismal Tide It will first eat anyone younger than Boomer.
After that carnage, we may pause |
|
Date: August 3rd, 2026 6:24 PM Author: law is ea$y (🧐) boomers are already dead man it's going to eat like 90% of xo aged workers (http://www.autoadmit.com/thread.php?thread_id=5888981&forum_id=2],#50043206) |