Like another comment says this is is not recent - these have been up on YouTube for a while. If you're in the bay area, these are free to attend (and recommended if you're a Knuth fan ofc), but reach early because the seats get filled up. Also at the end of the lecture Knuth gives away free copies of some of his books. These are usually in a pile, and are given first-come-first-serve; which is a polite way of saying there can be a mad rush for them :-)
I have been lucky to attend 3 of these in person (technically 4, but I had to leave one early), and more than the lectures themselves, Knuth's attention to detail and enthusiasm have been inspirational. He's 85 now - I would consider myself lucky if I have a third of his mental acuity at that age!
TAOCP Vol 4B is the most recent book in the TAOCP series which covers SAT solvers - apparently Knuth's deep dive in the area is recent, in due course of which he implemented a few SAT solvers (5, if i remember right). Talk about dedication. Since I have dabbled in using SAT solvers a bit, this is the volume I got autographed by him at the 2023 talk.
An interesting thing that came up was someone mentioned to him that a lot of results in 4B were empirical (including those around the topic he covered this time: dancing cells). His response was that theoretical analyses is still to catch up in these areas, so empirical comparison is the best you can do. I thought that was an interesting parallel to how ML has evolved.
IIRC... Slashdot was acquired by andover.net which was acquired by VA Linux which later changed its name to SourceForge.net, then was sold off to Dice.com then was sold off to the current owners.
While I love these videos, they have been online for several years now. The Slashdot headline is a misunderstanding of the article it is based on[1] which, when it says "Recently Stanford uploaded…", means a few years ago (before which too they were online, but not grouped into the same set of playlists on YouTube).
Anyway, the article this is based on[1] is a good summary of this year's talk. For several years now, this website (thenewstack.io), and specifically David Cassel, have been writing excellent summaries of each year's talk (see the "Previous Donald Knuth Christmas Lectures" at the end of the post, which links to everything from 2017 and later).
Reading a few of the written summaries, I’m finding them lacking. They often elide or even misstate important details and definitions; and the things the author focuses on are sometimes odd and make me think they aren’t able to understand the mathematics.
Whatever small faults there may be, each year these are the only reporting of these lectures at this level of detail. And the attention to humorous anecdotes or Knuthisms is not a problem; these are what those of us who attended the talk also remember. (The actual mathematics is often just a paragraph or two, or one problem and solution, in TAOCP and can be read there -- which is another indication of how "packed" the books are.)
I have been lucky to attend 3 of these in person (technically 4, but I had to leave one early), and more than the lectures themselves, Knuth's attention to detail and enthusiasm have been inspirational. He's 85 now - I would consider myself lucky if I have a third of his mental acuity at that age!
TAOCP Vol 4B is the most recent book in the TAOCP series which covers SAT solvers - apparently Knuth's deep dive in the area is recent, in due course of which he implemented a few SAT solvers (5, if i remember right). Talk about dedication. Since I have dabbled in using SAT solvers a bit, this is the volume I got autographed by him at the 2023 talk.
An interesting thing that came up was someone mentioned to him that a lot of results in 4B were empirical (including those around the topic he covered this time: dancing cells). His response was that theoretical analyses is still to catch up in these areas, so empirical comparison is the best you can do. I thought that was an interesting parallel to how ML has evolved.