Something (unrelated) that surprised me here is that the author has two decades of posts on this blog, leading back to their first post about a new version of C# (https://codeblog.jonskeet.uk/2005/09/)

Personally as a younger software guy I had no idea that C# had type inference back in 2005- an interesting perspective shift!

jon skeet is somewhat famous for topping the stackoverflow leaderboards since the beginning of time

https://stackoverflow.com/users?tab=Reputation&filter=all

- Users don’t mark Jon Skeet’s answers as accepted. The universe accepts them out of a sense of truth and justice.

- When Jon Skeet's code fails to compile, the compiler apologises.

- Jon Skeet does not use revision control software. None of his code has ever needed revision

Replace “Jon Skeet” with “Fabrice Bellard” and you’ve got yourself a real meme.

type inference in programming languages is more than 50 years old. ML had it in 1973.

jon skeet is one of the most famous programmers of all time. He might not be famous like Knuth is but he is everyday programmer famous.

Leaky abstractions.

What about it? The article isn't about leaky abstractions perse, it's about incorrect abstractions.

From the article: "Option 3: Embrace the leakiness"

As someone who wrote more abstractions than he cared for in the very similar domain of media technology, I thinkbthis is the right approach. In an ideal world the mixer abstraction shouldn't care about the type of hardware, but that kindnof media tech is far from creating an ideal world.