One major gripe I have is the infamous "TTFP"(Time to First Plot) problem is more or less solved, these improvements haven't manifested in any other areas.

Here is time to first read csv problem. I am on newest Intel Macbook Pro

  cat read_csv.jl
  using CSV
  using DataFrames
  df = CSV.read("hello.csv", DataFrame)
  println(df)

  cat hello.csv
  col1,col2,col3
  1,2,3
  4,5,6

  time julia read_csv.jl

  julia read_csv.jl  10.48s user 0.37s system 102% cpu 10.609 total