I'm a student at Stanford studying Physics with a specialty in Electro Mechanical Systems, and minoring in Computer Science in the deep learning track. My current focus at the moment is in working through Relativity and Quantum Mechanics while reading some of the canonical papers in machine learning and implementing them in Jupyter notebooks.
I'm also taking the time to learn Web Development (hence this website), read my thoughts on why I think it's still worth learning to code, so you should see my Github begin to slowly populate with some other projects as I learn more JS, along with my attempts at implementing papers.
I came late to my interest in STEM so I don't have many projects to show off (yet). Instead I can offer some of the essays I've written through my first year at Stanford and over the Summer, which, while admittedly less difficult to produce, showcase important aspects of my perspective on life and learning.
Close to 20% of students in Stanford's introductory CS courses have been failed for cheating in the past year, and these are only the one's who got caught (I know multiple who didn't). What are the consequences to human understanding when we outsource our work to AI? Are these skills becoming antiquated, and these students simply ahead of the curve, or are they throwing away the privelege to learn?
What does a future where people don't have to work look like? How will humans derive meaning in when there is such radical abundance? Are considerations of this future necessary to prepare for a super intelligence, or are they an unhelpful thought experiment, or worse, gross indulgence?
The 2009 Star Trek rebot featured a young Spock immersed in an AI powered tutoring session. Are Vulcan Pod's the future of education? Does Spock's struggle as a child and eventual departure of Vulcan go beyond him being half-human, and instead reflect a subliminal message about the need for social educations by the writer? Probably not, but for us humans on earth, it still begs the question: what will education look like when AI both knows the content and the student better than a human teacher ever could?
If there's one traedgy that's come out of the AI era it's that Andrej Karpathy didn't copyright the term "Vibe-Coding". God only knows how much money he'd have with a 1% royalty. However another term Karpathy is fond of, "leaky abstraction" has been on my mind. I can't help but think AI assisted coding, at least for industry grade software production, is a (very) leaky abstraction, one which still pays dividends to go under.