Video status, emerging from the rabbit hole, and England.

Hi Everyone,

First off, there will be a new probability video within the next week.  Thank you for your incredible patience on this front!  The topic, broadly speaking, is a case study for Bayesian updating in a continuous context, revolving around the beta distribution.  The script ended up long enough that I think it makes sense to split this one into two.  At the moment I'm on a trip to England (more on this below), so it seems best to wait on doing the voice recording until I'm back home.  In addition to those two videos, my current plan is to have a third simpler one out by the end of March.

As I mentioned in the last post, the majority of the past several weeks have gone to upgrading the animation library manim.  The main aims have been 1) increase the scope of what can be conveniently animated by reimplementing the back-end to be based on shaders, and 2) speed up the animation process by making it more interactive.  So in theory, at least if we look past research/writing and focus only on animation, videos this year should come more quickly and look better.

After many rabbit holes and lots of learning, I didn't quite get to everything I had hoped to on the list, but at some point, I had to just call it and jump back into making the next video.  On the shader front, things are all set up now.  For the next two projects, there won't actually be a need to do anything all that fancy visually; the explanations don't call for anything more than the usual vector graphics.  But I can certainly anticipate a few projects coming down the pipeline where this will make a difference.  In either case, behind the scenes, things render faster, but that's more noticeable to the author than the viewer.

On the interactivity front, although I only got minimally into the plans, as I've been animating this last video it definitely makes a big difference.  In short, there's no delay between typing a line of code and seeing the effect it has on the scene, and there are the humblest beginnings of an actual GUI intermixed with the coding when that kind of interaction is called for, which could be built out more as time goes on.

All this will hopefully be made more concrete when/if I put out some kind of "behind the scenes" video.  But I'm not sure that makes sense until the process is less in flux, and until there's been a decent outflow of actual math lessons on the channel.

England!

I came over to speak at Oxford this last week, which holds a special place in my heart ever since doing a term abroad there during college.  There had been a lot of potential energy welling up to come to England, so this invite was really just an excuse to finally make the trip.  

Among other things, Matt Parker and I filmed a collaboration together, which is on a particularly delightful prisoner puzzle.  We'll do the main puzzle on his channel, and some interesting math around showing why subtle changes to the puzzle make it impossible on 3blue1brown.

The week ended up including several other little bits of filming, too, including one little with a smaller channel called "Tom Rocks Math", run by the guy who invited me to Oxford, and some with Brady Haran's channel Objectivity at the Royal Society.

Matt Parker also runs something called "An Evening of Unnecessary Detail" with Steve Mould and Helen Arney, which is kind of like a comedy show targeted at nerds.  He invited me to do a 10-minute bit for it while I was here, which is certainly very different from my usual comfort zone, but it ended up being incredibly fun.  It was recorded, and as soon as I get the video I'll share it with you.  The same goes for all these recordings, of course.

-Grant