tutorials on how to do "arithmetic on machines with just binary add, subtract, logical and shift operators, making no use of special hardware support for such complex operations as BCD arithmetic or multiplication and division" very good, very helpful
nicely explained
"Embedding tricks and chain-of-thought prompting simply extends their ability to do more sophisticated pattern matching. The mathematical results imply that you can always find compositional tasks whose complexity lies beyond a given system’s abilities."
i always lose track of this site, which is very nice
"Noise is not a complicated concept, but it has many subtleties." practical, detailed introduction with code samples
3Blue1Brown calculus lessons are back! in pog form. i mean blog form
presents "a different approach concentrating on the migration from Python to Numpy through vectorization..." with "a lot of techniques that you don't find in books and such techniques are mostly learned through experience"
rhetoric and conic sections. this is amazing
like k-means but 1d, good overview
"a collection of texts and images concerned with digital curvature. It seeks to understand when a curve starts to be tangible, or what might give a sense that they can or can't be handled. What makes a curve smooth? When does it seem appropriate to interrupt it?"