Hey, I'm Owen.
I think rationality is pretty rad.
Whoa! I wrote about something similar here a while ago under the same name, at least about the aesthetics part.
Note: Anna Salamon has a public response on the FB post here (unsure to what extent it's official)
Seconded. In my view, the anecdotes are there such that the idea is more salient and hangs around longer in your head.
Sure, you can read 10 self-help summaries in an hour, but I don't think that gives you 10x the same amount of benefit as reading about one concept for an hour. (If anything, I don't even think you get 1x the same amount of benefit, as you have to factor in potential confusion sorting everything out, etc.)
The padding can also be useful if you're trying to learn via example, or learn what the stereotype of The Concept looks like.
LFD was my first intro to statistical learning theory, and I think it's pretty clear. It doesn't cover the No Free Lunch Theorem or Uniform Convergence, though, so your review actually got me wanting to read UML. I think that if you're already getting the rigor from UML, you probably won't get too much out of LFD.
I'm curious if you've looked at Learning From Data by Abu-Mostafa, Magdon-Ismail, and Lin? (There's also a lecture series from CalTech based off the book.)
I haven't read Understanding Machine Learning, but it does seem to be an even more technical, given my skimming of your notes. However, the Mostafa et al book does give a proof of why you can expect the VC dimension to be polynomially bounded for a set of points greater than the break point (if the VC dimension is finite), as well as a full proof of the VC Bound in the appendix.
Hmmm. I agree with you that fingernail biting didn't seem to fit the paradigm. However, I did Google "stop biting fingernails", though, to see if there was any domain specific suggestions. (You may have already done this.)
Two things that maybe seemed promising:
Something else which seems maybe useful is to be mindful/reflective after you've noticed that you've done it.
Otherwise, I (at least right now) don't know much about breaking habits without knowing the trigger.
Thanks for the info, Ozzie!
I checked out Observable some more. I think it might actually be a little heavier then what I want. Unsure if I'll do the coding exercises beforehand (and just post the results + code), or if I'll go through the work of setting up an interactive notebook so readers can follow along.
I looked into self-hosting it because it seems the default option is creating a notebook hosted on their site. My understanding is that there's a way to embed notebooks onto my own sites (or the runtime environment is open-sourced?)
I'm going to spend some of the winter holidays working on Abu-Mostafa et al's Learning From Data's problem set. I think this should be fun, and I'll also look into learning Observable for some interactive notebooks for the coding problems.
This piece was helpful in outlining how different people in the AI safety space disagree, and what the issues with Paul's approaches seem to be. Paul's analogies with solving hard problems was especially interesting to me (the point where most problems don't seem to occupy a position midway between totally impossible and solvable). The inline comments by Paul were also good to read as counterpoints to Eliezer's responses.
Sidenote: Loved the small Avatar reference in the picture of the cabbage vendor.