Python vs. Julia ODE Solver

This section alludes to it although maybe we should say it more explicitly.
https://docs.julialang.org/en/v1/manual/performance-tips/#man-performance-abstract-container

julia> [1 1.0 "sfsdf"]
1×3 Array{Any,2}:
 1  1.0  "sfsdf"

julia> typeof(ans)
Array{Any,2}

julia> (1, 1.0, "sfsdf")
(1, 1.0, "sfsdf")

julia> typeof(ans)
Tuple{Int64,Float64,String}

An array of Any is abstract and is hard for the compiler to write optimized code for. See @stevengj 's lecture to understand the problem there: