First Impressions of Spring 2026
The Spring 2026 semester is off to a start. So far, it's been pretty chill, but I think it's going to become horrible in a couple weeks. Here are some first impressions of my classes, with the ones that I like closer to the top, and my less favorite ones towards the bottom.
So far, I am really enjoying this class. It's like a combination of Math 104 and things from 270. It's honestly made me realize that I like 104 more than I realized at the time. The problem sets are not impossible, but they are fun. The professor, Venkat Anantharam, is a pretty good lecturer; he messes up some things sometimes, but he sometimes drops nuggets of wisdom that blow my mind. I took this class because I really liked the duality and optimization part of 270, and so far, this class is rivaling 270 in being the best class I've taken at Berkeley. I think it's definitely easier to understand than 270; I think I understand (in percent of lecture understood) it an order of magnitude better than 270. I do hope that we use more stuff from 104, as that's brought me the most joy, and I'd like the homework to have analysis as well. The homeworks are pretty good; they aren't insanely hard, but some problems need some neat tricks.
Lectures for this class are very fun. The class is taught by Tony Yen, who is a big guy at ASML. He is a good lecturer; he's funny and has some very good moments. I've never had a professor pretending to be a German boss at Zeiss punch a projector screen, cursing in German, complaining about the numerical aperature. His stories are very fun to listen to, and the class itself is pretty relaxed. Homeworks are due every two weeks, and mostly cover optics and extensions/proofs of things discussed in class. I feel like this class has made me realize that optics and E&M is pretty interesting. I remember that I was bored of E&M last year but this class has made me like it again. Highly recommend.
My biggest complaint about this class is that it's at 8 AM. Sergey Levine is one of the best lecturers I've had; his slides are excellent, and it's rare that you have a teacher that's both amazing at research and at teaching. Despite that, I've found it hard to attend most of the lectures; the lecture hall is really far at a very early hour. The homeworks aren't diabolical; it's mostly coding algorithms that were taught in class. I think I'd prefer some more theoretical problems in the homework; that may change in the coming weeks. Excellent class.
This class is both all that they say and not all that, at least so far. Obviously, projects and homeworks are some of the longest I've seen in any class, but I feel like they give enough time for it to be manageable. Obviously, this might change when they release the later project.s The topics of the class are pretty cool; it's honestly incredible how much complexity there is in running something as simple as a hello world program. The codebase is large, and I'd say that I'm not a fan of the GDB, but I suppose that those are necessary evils. I did get cooked by the first midterm, so I think I'll have to devote more time to this class. Kubi's lectures are interesting, but sadly, I don't really go to them. Overall, a pretty good class.
This class is far too early for me, at 8 AM. I'm not really sure what's going on in the class. From the one lecture I've been to, the professor is pretty good. However, I feel like the homework is a bit excessive; every week is pretty diabolical, and I'd prefer problem sets that are shorter but trickier. I came into the class late, because I was hoping to get into another section (one at a more reasonable hour), so I can't say much for this class either. I hope to see some beauty in this course, and I hope I can use it in physics.
I don't know what to say about this class. I was excited to take it, but ever since the curriculum changed, it feels like the class has lost a lot of its allure. It's very applied, and it doesn't feel like it covers much. The infamous homework is gone; it's been replaced by long coding homeworks. While I suppose it is important to provide some coding in the homework, the lack of thought-provoking theoretical questions really hurts learning. I also did not like that the first homework used pandas. I feel like 189 should provide a better mathematical foundation for later classes in ML/AI, and it feels like it's not provided that. I haven't gone to any lecture after the first, so I can't say anything for the lectures, but judging from the homeworks, it seems like the class has fallen off a cliff (in my opinion).
I hope you enjoyed reading this.
Sid