Having created a lot of content for JuliaActuary.org, I spent a lot of time over the past few years working with my co-author Yun Tien Lee creating a book about bringing programming and related skills to actuaries and other financial professionals. We released the book earlier this year and it’s available completely free online or as a paid print edition.
I think it’s a great time to learn topics like this, because since we wrote the content for the book AI has become a big force in technical disciplines, and having the knowledge and confidence to interact with an AI is (IMO) becoming an essential skill. AI isn’t good enough to blindly trust it, and you need to have the foundational knowledge to suss out the good from the bad. The book itself spends virtually no time on AI and instead focuses on the fundamentals.
Feedback thus far has been great and any constructive feedback is still welcome!
Just curious… What is the deal with Julia? For most people I talk to, the debate about most popular programming language is R vs Python. What does Julia bring to the party?
Julia is high level, general purpose, and interactive like Python, but way more performant: Benchmarks – JuliaActuary
Julia brings some true innovative features like Multiple Dispatch, language wide automatic differentiation, and incremental compilation.
Julia has a lot of modern features you want from a language and related tooling (package manager, environments, etc)
Julia’s ecosystem covers almost all of the relevant packages you might want vs Python (a good Chainladder package being a notable exception).
Julia’s just nice to work with and it scratches many itches as it can be object oriented, functional, array-oriented, etc. You can match the paradigm to best fit the problem rather than trying to make everything some OOP variant.
Also, in the age of AI it seems like you can optimize for the best underlying programming language rather than the one with the most inertia (e.g. the argument in this post). I think it’s important for a human to review so a highly readable language like Julia works well (as opposed to say, having the AI write in Fortran which is highly performant but hard to read).
I spent several years writing R, later Python as my primary language. My preference for Julia solidified when I re-wrote a reinsurance model from Python to Julia ~6 years ago now. Python is a wonderful scripting language but no longer my preference for heavy computational work.