Mostly, a great talk. It was informative, and I felt like I really understood what you were doing! That was really cool! You probably want a laser pointer for the real talk to point at the slides.
Some nit-picky things (almost entirely slide-formatting things; the presentation itself was solid):
The slides for introducing the initial language were incredibly crowded. I would remove the judgment rules and the distribution pictures; simply provide the implementation and invocations and describe the behavior.
Add more space between the 'slide title' and the 'section title'; make it clear what we should really be paying attention to.
Why \x . cx instead of just c ?
Casually dropping the word 'continuation' to explain an integral is... confusing. Ah, the second slide clarified. I'd try to be stronger about that connection.
Please use && or and instead of ^ for and.
The generalization of the density examples was a bit hard to follow. More directly (maybe formally) stating the induction and the cases would provide nice sign-posting.
What are the emojis for? They clutter the slide; maybe strip them?
I missed why we went to lists, or what the lists actually contain. Presumably the density; how do we interpret that result? Also, don't use Haskell's list comprehension syntax there because it's easy to confuse [ .. | .. ] from [] on a slide.
In The Paper could be 3 slides, and broken out more. Also, actually hearing about applications would be swell.
probably good idea to avoid more than one title for one slide
too many information on the "Calculating probability densities" slide, it was somewhat hard for me to follow, also in the "Deriving density calculator".
it may help to be consistent on showing lambda functions (with or without black background)
it's unclear from the slide whether the Exp e | Var v | Lev v e e ... are already in the paper, or are those future work..
Mostly, a great talk. It was informative, and I felt like I really understood what you were doing! That was really cool! You probably want a laser pointer for the real talk to point at the slides.
Some nit-picky things (almost entirely slide-formatting things; the presentation itself was solid):
&&
orand
instead of^
for and.[ .. | .. ]
from[]
on a slide.In The Paper
could be 3 slides, and broken out more. Also, actually hearing about applications would be swell.