|
Torsten Scholak
@
tscholak
Montréal
|
|
applied research scientist @element_ai, formerly physicist @uoft, into natural language processing, Bayesian stats, and functional programming in @haskell_lang
|
|
|
2.297
Tweetovi
|
2.421
Pratim
|
923
Osobe koje vas prate
|
| Tweetovi |
|
|
Torsten Scholak
@tscholak
|
4 h |
|
Seems like that this insight is the information that was conveyed by the movie, so I guess it’s a contradiction?
|
||
|
||
|
|
Torsten Scholak
@tscholak
|
13 h |
|
thanks for getting back to me. I’m indeed curious about these fundamental reasons you mentioned. I looked at the blog post again, and it is not more clear to me what these could be, except that @dhall_lang’s python bindings are still a work in progress, github.com/SupraSummus/dh…
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
15 h |
|
I wished I could easily use @dhall_lang for this
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
5. velj |
|
come work with @austinvhuang, @junjihashimoto3, @apaszke, @SamStites and me twitter.com/austinvhuang/s…
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
3. velj |
|
unpossible!
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
2. velj |
|
I can relate. Whenever my wife @TewfikianEcon and I are in Belgium, we buy plenty of Dessert 58 bars, and whenever we are in Germany, we are shopping for Lindt bars that they don’t sell in Canada.
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
2. velj |
|
I think what confused me was that I lacked a good working definition of “expressions”, “statements”, and “terms” that would highlight their differences.
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
2. velj |
|
thanks @fried_brice! I googled “expressions vs statements”, and found this: stackoverflow.com/questions/4728…. I understand now. thanks also to @fresheyeball for the IO () example.
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
2. velj |
|
I’m using haskell, and I don’t know what this means, what exactly is claimed by this statement, and why it makes sense :/ — what is expression-oriented vs statement-oriented? and why is the former safer than the latter?
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
2. velj |
|
Very interesting indeed
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
31. sij |
|
Congratulations, looking forward to your contributions!
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
31. sij |
|
especially since you invented Haskell’s outside-in constraint solving algorithm :D — coincidentally, I started reading on it yesterday after musing over @Lowert’s thesis on matchability and partially applied type families...
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
31. sij |
|
the bollard would have been fine if it had worn a shiny west and a helmet
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
31. sij |
|
Sounds interesting, because I’m active in that space myself, working towards joint approaches to neural program synthesis, constraint solving, and type inference... but asking for compiler experts exclusively in C, C++, and Python doesn’t make sense IMO. Why not Haskell?
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
30. sij |
|
the n-ary version of this could be achieved with trchniques introduced by @haskell_cat in his n-ary-functor library twitter.com/Jose_A_Alonso/…
|
||
|
|
||
| Torsten Scholak proslijedio/la je tweet | ||
|
Austin Huang
@austinvhuang
|
30. sij |
|
Timely post on Jax’s vmap in Haskell by @ezyang. Will have to give these ideas a try in @hasktorch! twitter.com/Jose_A_Alonso/…
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
28. sij |
|
You could use a library that parameterizes tensor types over compute devices, twitter.com/tscholak/statu… twitter.com/tscholak/statu…
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
28. sij |
|
I’m jealous :D
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
28. sij |
|
slippery road, Rob. soon you will discover automatic differentiation Hamiltonian Monte Carlo and NUTS...
|
||
|
|
||
|
|
Torsten Scholak
@tscholak
|
27. sij |
|
use @dhall_lang
|
||
|
|
||