@lgodowsky.bsky.social
Or else, he would deport her
January 16, 2026 at 2:50 AM
Or else, he would deport her
Prez, Could you exchange a FIFA Peace Prize for a Nobel Peace Prize?
January 10, 2026 at 2:29 PM
Prez, Could you exchange a FIFA Peace Prize for a Nobel Peace Prize?
He forgot to say “Oh BTW, we had the CIA bribe Venezuela’s military leadership. And its VP was willing to turn in Maduro.”
archive.is/qVXAp
archive.is/qVXAp
archive.is
January 5, 2026 at 8:20 PM
He forgot to say “Oh BTW, we had the CIA bribe Venezuela’s military leadership. And its VP was willing to turn in Maduro.”
archive.is/qVXAp
archive.is/qVXAp
Personal experience also tells me that LLMs cannot generate even two lines of MIPS code correctly. Same problem with elliptic curve encryption.
December 23, 2025 at 4:16 PM
Personal experience also tells me that LLMs cannot generate even two lines of MIPS code correctly. Same problem with elliptic curve encryption.
No solutions to your problem without human’s expertise due to Rice’s theorem
December 23, 2025 at 3:57 PM
No solutions to your problem without human’s expertise due to Rice’s theorem
Yes, but correctness is still relative to some formal specification of a requirement. LLMs have no internal mechanism of constructing these formal requirements, much less formal verification.
December 23, 2025 at 3:47 PM
Yes, but correctness is still relative to some formal specification of a requirement. LLMs have no internal mechanism of constructing these formal requirements, much less formal verification.
Looking forward to just scanning thru the preprint!
December 23, 2025 at 2:43 PM
Looking forward to just scanning thru the preprint!
Yes. To be precise, LLMs sequentially generate the next most likely token based on the prompt and the datasets they were trained on. This is their metrics, i.e., merely the syntax captured by the tokens. No formal semantics of the language involved. Rather, the process is a brute-force method.
December 23, 2025 at 2:21 PM
Yes. To be precise, LLMs sequentially generate the next most likely token based on the prompt and the datasets they were trained on. This is their metrics, i.e., merely the syntax captured by the tokens. No formal semantics of the language involved. Rather, the process is a brute-force method.
He doesn’t understand the implication of “For any input distribution D_n, …” He wants you to prove the theorem for a “specific” distribution 😀.
December 21, 2025 at 11:33 PM
He doesn’t understand the implication of “For any input distribution D_n, …” He wants you to prove the theorem for a “specific” distribution 😀.
Quite a radical ideal
December 21, 2025 at 8:47 PM
Quite a radical ideal
Replace “George Soros” with “King Trump” and you get an entirely interpretation
December 21, 2025 at 8:28 PM
Replace “George Soros” with “King Trump” and you get an entirely interpretation
The field has mostly been a pseudoscience from its inception.
November 6, 2025 at 9:49 PM
The field has mostly been a pseudoscience from its inception.
Practicing before his really dances in the Epstein Ballroom
November 6, 2025 at 2:10 PM
Practicing before his really dances in the Epstein Ballroom
The chain complex definition for HAM, SAT, or general computational problems is incorrect. What is the module of coefficients? It wouldn't work over Z. So it makes no sense to formulate homology groups. Homotopy type theory plays zero roles here. Conclusion: another AI slop.
October 26, 2025 at 12:51 PM
The chain complex definition for HAM, SAT, or general computational problems is incorrect. What is the module of coefficients? It wouldn't work over Z. So it makes no sense to formulate homology groups. Homotopy type theory plays zero roles here. Conclusion: another AI slop.
😂 you guys should have learned how SNAP works at the state level. Poorer red states would be more miserable.
www.milkenreview.org/articles/red...
www.milkenreview.org/articles/red...
Red State, Blue State
www.milkenreview.org
October 23, 2025 at 2:31 PM
😂 you guys should have learned how SNAP works at the state level. Poorer red states would be more miserable.
www.milkenreview.org/articles/red...
www.milkenreview.org/articles/red...
As an ignorant theoretical physicist, he doesn’t even know that Roger Penrose’s twistor theory is based on classical algebraic geometry.
October 22, 2025 at 8:59 PM
As an ignorant theoretical physicist, he doesn’t even know that Roger Penrose’s twistor theory is based on classical algebraic geometry.
Be open-minded and learn advanced math. Classical algebraic geometry helped Penrose formulate his twistor theory.
pirsa.org/c24016
pirsa.org/c24016
Higher Categorical Tools for Quantum Phases of Matter | PIRSA
Quantum phases have become a staple of modern physics, thanks to their appearance in fields as diverse as condensed matter physics, quantum field theory, quantum information processing, and topology. The description of quantum phases of matter requires novel mathematical tools that lie beyond the old symmetry breaking perspective on phases. Techniques from topological field theory, homotopy theory, and (higher) category theory show great potential for advancing our understanding of the characterization and classification of quantum phases. The goal of this workshop is to bring together experts from across mathematics and physics to discuss recent breakthroughs in these mathematical tools and their application to physical problems.
pirsa.org
October 21, 2025 at 3:41 PM
Be open-minded and learn advanced math. Classical algebraic geometry helped Penrose formulate his twistor theory.
pirsa.org/c24016
pirsa.org/c24016
Mathematical proofs cannot automatically be constructed by ANY system we've today and even in the foreseeable future. This is why we have proof assistants in which humans still have to do heavy liftings. Tell us how you can encode all the traditional mathematical axioms in a formal system.
October 17, 2025 at 10:58 PM
Mathematical proofs cannot automatically be constructed by ANY system we've today and even in the foreseeable future. This is why we have proof assistants in which humans still have to do heavy liftings. Tell us how you can encode all the traditional mathematical axioms in a formal system.
An implementation of Terry Tao's idea: arxiv.org/abs/1707.07807
On the universality of the incompressible Euler equation on compact manifolds
The incompressible Euler equations on a compact Riemannian manifold $(M,g)$ take the form \begin{align*} \partial_t u + \nabla_u u &= - \mathrm{grad}_g p \mathrm{div}_g u &= 0. \end{align*} We show th...
arxiv.org
October 17, 2025 at 9:04 PM
An implementation of Terry Tao's idea: arxiv.org/abs/1707.07807
Acemoglu’s claim is bogus. Did he study the history of computing, for example?
October 17, 2025 at 1:49 PM
Acemoglu’s claim is bogus. Did he study the history of computing, for example?