The most prestigious law school admissions discussion board in the world.
|
 |
 
|
CEO of Anthropic majored in Jewish Physics at Stttanford
Poast new message in this thread
Date: April 28th, 2026 7:50 AM Author: Topaz well-lubricated mexican ticket booth not flame (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848490) |
Date: April 28th, 2026 7:57 AM Author: Bat shit crazy contagious love of her life ratface We talk cosmology after every minyan, bro. Every Rabbi is out there learning C* algebra thanks to ARE newest greatest Jew — this generation’s Einstein — Jacob Barandes. Though technically another Jew was first to use this maths — bro named Segal. Basically - we invented everything of note. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848498) |
 |
Date: April 28th, 2026 7:58 AM Author: Topaz well-lubricated mexican ticket booth Tech bros should all be studying Jewish Physics now that AI can write software (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848500) |
|
Date: April 28th, 2026 8:01 AM Author: Bat shit crazy contagious love of her life ratface Eric Weinstein thinks humans have no more role in math or physics anymore and that AI is going to solve everything in 5-7 years and is already doing original research with human prompts and that it’s already a golden age. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848505) |
|
Date: April 28th, 2026 8:01 AM Author: Topaz well-lubricated mexican ticket booth He's correct lol. Still requires an esoteric math/physics priesthood to verify everything. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848508) |
|
Date: April 28th, 2026 8:05 AM Author: Bat shit crazy contagious love of her life ratface Biggest lesson from AI so far is that its showing proofs are thought were proofed and done are wrong and that humans are so bad at math that entire generations have just been wrong about centuries old problems. So far it’s torn apart more maths than it’s solved…. Not sure humans can verify shit…. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848514) |
|
Date: April 28th, 2026 8:10 AM Author: Topaz well-lubricated mexican ticket booth Yeah good point. "everything solved in 5-7 years" is way too strong though. Not everything will get solved but it will be significantly verified. And you are right that possibly the verification will get automated too without need for human input. Agree it has gotten to the point where its better than only the very top experts in Math and exceeds it in many respects. Some of the stuff that people consider "wrong" in its proofs too is more about readability. A lot of human math, the language is optimized for social understandability and pedagogy but AI often finds structures that should be "correct" but humans don't do it that way. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848523) |
|
Date: April 28th, 2026 8:16 AM Author: Bat shit crazy contagious love of her life ratface Weinstein is intentionally flaming for the clicks and followers. The reality is we have more maths than we know what to do with and the problem since the 40s is having the experiments catch up with the 500 theories that appear to work on paper but can’t be verified. Every experiment now takes a billion bucks and 3-10 years and I’m not even talking about the particle accelerators. Radioinferometry is the low hanging fruit and it’s still $400 million to get one satellite built. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848533) |
|
Date: April 28th, 2026 8:29 AM Author: Topaz well-lubricated mexican ticket booth Yeah there are tons of pure maths and mathematical physics and new stuff still coming out like HoTT and Condensed Mathematics etc. A lot of it probably restatements of known results in different language noone noticed because too much stuff and too many subdisciplines. But we still probably need more to actually solve the deepest problems in physics. I think basically all of it will need to be "refactored" and debugged by Mythos or better AI. Instruments and sensors are catching up but probably still 20-30 years away for confirming a lot of theoretical stuff. Some things like planck scale might not even be able to be experimented on in principle at least not using known physics. I think you are right that radioinferometry in space and also I think gravitational wave detectors will be the low hanging fruit coming up (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848548) |
|
Date: April 28th, 2026 8:03 AM Author: Bat shit crazy contagious love of her life ratface AI will need us for our opposable thumbs to build experimental hardware for a little while longer so the experimental physicists will still have a job for a bit but the guys just doing math are already fucked. (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848510) |
|
Date: April 28th, 2026 8:11 AM Author: Topaz well-lubricated mexican ticket booth Still need humans to talk about how "beautiful" and "deep" the equations are though (http://www.autoadmit.com/thread.php?thread_id=5861273&forum_id=2)#49848525) |