Jump to content

General Relativity: Difference between revisions

General Relativity (edit)

Revision as of 07:39, 14 December 2025

6,951 bytes added ,  14 December 2025
→‎2024
Line 1,881: Line 1,881:
}}
}}
|timestamp=9:14 PM ¡ Sep 13, 2024
|timestamp=9:14 PM ¡ Sep 13, 2024
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1873084952444846239
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=[Note to astute readers, “figured out” is doing some lifting here. Not every post needs a footnote system however.]
|thread=
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1873076995376243129
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=Such a strange way to refer to Europeans. You know, that tribe who figured out even tempered harmony, movable type, the double helix, the grasshopper escapement, and [[General Relativity|general relativity]].
I care. And I can’t begin to tell you how much. We should really pay attention to that tribe.
|quote=
{{Tweet
|image=EdKrassen-profile-4iMzq_Tk.jpg
|nameurl=https://x.com/EdKrassen/status/1873049109663457390
|name=Ed Krassenstein
|usernameurl=https://x.com/EdKrassen
|username=EdKrassen
|content=White people will one day no longer exist. It’s just a matter of time. Who cares?
|timestamp=4:51 PM ¡ Dec 28, 2024
}}
|timestamp=6:42 PM ¡ Dec 28, 2024
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1873077786585559382
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=Ed, I don’t know you, but put “Indians” or “Zulus” or “The Inuit” where you have “White people” (I.e. European Ancestry) and you will likely see the problem in what you wrote.
|timestamp=6:45 PM ¡ Dec 28, 2024
}}
|timestamp=7:13 PM ¡ Dec 28, 2024
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1894046135242858876
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=Final thoughts: all the LLMs choke on these problems. This isn’t so far about Grok. It’s about what competence is actually needed for super human intelligence to make progress in physics.
We haven’t even gotten started on real math with these models. We need to go way way harder on pushing them before we talk about using them to understand the universe.
Happy to continue via dm or voice. Up to you. Good luck and Keep Going!!
|thread=
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1893871484814114962
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=It’s very impressive. But this is wildly misleading.
I give it completely straight forward problems in Linear algebra/reoresentation theory and it refuses to stop guessing answers. It “bullshits”.
It is not remotely close to passing a mathematical Turing test in THIS Iteration.
|quote=
{{Tweet
|image=MarioNawfal-profile.jpg
|nameurl=https://x.com/MarioNawfal/status/1893779845177278746
|name=Mario Nawfal
|usernameurl=https://x.com/MarioNawfal
|username=MarioNawfal
|content=🚨GROK 3 GOES SUPERHUMAN—SOLVES UNSOLVABLE PUTNAM PROBLEM
None of the top 500 Putnam competitors fully solved this brutal math problem.
Grok 3? Crushed it in ~8 minutes.
Grok 3 isn’t just catching up, it’s surpassing human limits.
Source: @luismbat
|media1=MarioNawfal-X-post-1893779845177278746-GkgOmfoXgAA5zdm.png
|media2=MarioNawfal-X-post-1893779845177278746-GkgOnU0XQAEHx20.png
|timestamp=9:47 PM ¡ Feb 23, 2025
}}
|timestamp=3:52 AM ¡ Feb 24, 2025
}}
{{Tweet
|image=elon-profile.jpg
|nameurl=https://x.com/elonmusk/status/1893977704694755471
|name=Elon Musk
|usernameurl=https://x.com/elonmusk
|username=elonmusk
|content=Please cite examples
|timestamp=10:54 AM ¡ Feb 24, 2025
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1894040200319508954
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=Sure thing. It gets the following question reliably wrong but appears to correctly interpret the shorthand of what is being asked:
Q: Give the highest weights and dimensions of the Irreps that occur as direct summands of $\Lambda^2 \otimes \Lambda^2$ for SO(16) with Dynkin diagrams D_8.
|timestamp=3:02 PM ¡ Feb 24, 2025
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1894041322647924892
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=What is interesting is asking it why it failed despite numerous hints meant to help it. It appears to introspect correctly:
|timestamp=3:06 PM ¡ Feb 24, 2025
|media1=ERW-X-post-1894041322647924892-Gkj8iviaIAAbuMf.jpg
|media2=ERW-X-post-1894041322647924892-Gkj8ivkbEAAqyeW.jpg
|media3=ERW-X-post-1894041322647924892-Gkj8iwUWoAA7fzk.jpg
|media4=ERW-X-post-1894041322647924892-Gkj8iwUXoAAqMpd.jpg
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1894042305587229179
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=But it can’t make use of this capacity.
Not to tell you your business, but it appeared it needed to better know when it could use the attention heads to autocomplete and when it needed them to write an outside call to a reliable method in a trusted library.
|timestamp=3:10 PM ¡ Feb 24, 2025
|media1=ERW-X-post-1894042305587229179-Gkj9b_VWEAAcKPV.jpg
|media2=ERW-X-post-1894042305587229179-Gkj9b_hWsAAszfE.jpg
|media3=ERW-X-post-1894042305587229179-Gkj9b_XXIAA1TGz.jpg
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1894043561177571440
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=The question is actually an example of the kind of thing it needs to know to do any deep physics because both [[General Relativity|general relativity]] and the [[Standard Model]] are based around field strength as measured by curvature tensors. If you take one Lambda^2 to be differential forms along spacetime and the other Lambda^2 to be the Lie Algebra of the Lorentz group then the space of Lie algebra valued 2-Forms is exactly where the Riemann curvature decomposes into weyl, Traceless Ricci and Ricci scalar pieces.
|timestamp=3:15 PM ¡ Feb 24, 2025
}}
{{Tweet
|image=Eric profile picture.jpg
|nameurl=https://x.com/EricRWeinstein/status/1894044784106586488
|name=Eric Weinstein
|usernameurl=https://x.com/EricRWeinstein
|username=EricRWeinstein
|content=But because the other components never get occupied, Grok cannot find enough literature on the subject. There are always three other component Irreps above dimension 8 (where the number is 4) but there mostly folklore in mathematics and are thus not much written about. Plus pushing the calculation to R^{16} gets you away from the more prolific lower dimensional literature as well.
|timestamp=3:20 PM ¡ Feb 24, 2025
}}
|timestamp=3:26 PM ¡ Feb 24, 2025
}}
}}


Retrieved from "https://theportal.wiki/wiki/Special:MobileDiff/25077"