Finished reading Burkov, The Hundred-Page Machine Learning Book.
The book covers a lot of material at a perhaps problematic level of concision. It uses equations where other authors would use diagrams. It's quite good for review, but if you haven't encountered a topic before, you probably won't be able to follow the author's explanation. This could be useful as a textbook for a course where the instructor provides the necessary scaffolding to follow the text.