Read on Twitter
It's a shame that Stanford's computer science curriculum doesn't emphasize linear algebra. Linear algebra isn't a required course, not even for the AI track! (1/)
The main reason Stanford CS majors find AI classes hard is that for many of them, their linear algebra education started and stopped with putting matrices in RREF, computing determinants of 2-by-2 systems, computing eigenvalues of tiny matrices --- all by hand of course ...
Many (most?) CS students don't understand matrices as representing linear maps, have never heard of the SVD, aren't aware that derivatives are linear maps (leading to tons of confusion re: gradients) ... But it's not their fault!
In office hours for CS221, a student of mine actually became angry when she was struggling with the linear algebra on the homework, because she realized that the dept hadn't prepared her for it.
Instructors/TAs had to spend a lot of time on remedial linear algebra (computing gradients on the whiteboard by hand, explaining the chain rule, etc), distracting from the content of the course.
(In fact it was clear that some instructors/TAs themselves weren't educated in linear algebra ... you can imagine how that went)
The lack of emphasis on linear algebra is all the more surprising considering that the need to solve linear systems was one of the main reasons computers were built
(Stanford CS does half-heartedly recommend MATH 51, which nominally covers some linear algebra ... but to describe all the ways in which that class failed its students would require another thread.)
In the spirit of being constructive, here are some resources on linear algebra, ordered roughly from beginning to intermediate:
https://www.youtube.com/playlist?list=PLZHQObOWTQDPD3MizzM2xVFitgF8hE_ab
http://vmls-book.stanford.edu/
https://linear.axler.net/
https://people.maths.ox.ac.uk/trefethen/text.html
https://www.youtube.com/playlist?list=PLZHQObOWTQDPD3MizzM2xVFitgF8hE_ab
http://vmls-book.stanford.edu/
https://linear.axler.net/
https://people.maths.ox.ac.uk/trefethen/text.html
This thread was a lot more popular than expected. Just wanted to say that I loved my Stanford experience (BS/MS CS, now PhD EE). The CS programs have lots going for them (great systems & theory classes, accessible professors, etc) They’d just be even better with more lin alg 
For Stanford students, I’d recommend the following classes on linear algebra: ENGR 108, EE 263, and (if you like theory) MATH 113.
