10/07/2025

 wow ooops like two months have passed already. i have been trying to read and learn things but it's been a bit difficult to juggle everything going on in my life T_T

so the nobel prize winners are being announced recently which is very cool! for physics John Clarke, Michel Devoret, and John Martinis won for their experiments in quantum computing (which were apparently conducted in 1984 and 1985 which is a while back~). they showed that the phenomenon of quantum tunneling (passing through an energy barrier that wouldn't have been possible by classical mechanics, has applications in nuclear fusion and radioactive decay) can be observed on a macroscopic scale by building a circuit with superconductors which have electrons joined together in pairs called Cooper pairs. in physiology and medicine, Mary E. Brunkow, Fred Ramsdell and Shimon Sakaguchi won the prize for their research about peripheral immune tolerance that prevents the immune system from harming the body (?), discovering a new class of T cells. they figured out that a rare autoimmune disease in humans (IPEX) was caused by a gene mutation similar to a mutation they investigated in mice (scurfy mice) on the X chromosome, called the FOXP3 gene. later it was discovered that this gene controls the development of regulatory T cells. these discoveries are driving scientific research in finding potential new treatments to diseases!

persona vectors in chatbots: welp i bookmarked this a while ago. anyways these llms can adapt certain personas like evil, sycophantic etc; anthropic tests persona vectors by injecting these vectors artificially through steering and they do really seem to exist. this discovery allows us to monitor personality shifts during deployment which is helpful to both deployers and users, reduce undesirable personality shifts during training on the datasets that purposefully induce certain personas by using these vectors to intervene, and flag these datasets that might have their intentions hidden.

reviving enumerative geometry problems with "motivic homotopy theory"?? and then calculating quadratic forms in different number systems to help with solving these counting problems, removing the abstractness which makes it easier for younger students to get involved with this type of research :O.

interesting interview about ai and math intersection: we don't really know what problems ai can solve in math, and we shouldn't really view it as an "opponent." in fact, Dr. Ono says he uses ai as a tool in his research; "A good partner is one that is going to give you clues to conjectures that you can then either prove or build a body of work from." furthermore, ai is not as good as humans in asking questions.

now onto books and films. in school we read Changing Planes and it was very interesting imagining life on strange planets while you're waiting at an airport. personally i hate airports for being so bland so this book was an interesting concept. we watched some episodes of Watchmen (2019) and i really think it's interesting how the directors strung together all these scenes to create a social and cultural commentary on race, love, gender, history, superheroes, etc etc. 

a book that i read recently was called The Lantern of Lost Memories and it was about a photobooth we go to when we are between death and the afterlife. 3 characters and the photographer reminisce on the memories of their life that are all somewhat interconnected, and it was a really beautiful little book. i hope when i reach the photobooth i have a lot of good memories to think about :)

yeah in the past two months i've read quite a few books and watched/listened some interesting things lol. current music obsession is dream perfect regime, literally every mv/song is a piece of art

anyways that's it for now. :D maybe posting frequency will increase after this month...