Jacob Lifshay

> [@programmerjake](https://github.com/programmerjake) I'm going to leave a call like that to cryptographers like [@traviscross](https://github.com/traviscross). My understanding is that ChaCha has had more scrutiny. yes, but if it's specifically insecure, why...

> My only worry with Squares is that its quality depends on the key: > > > The key should be an irregular bit pattern with roughly half ones and...

> Before worrying too much we should actually measure how many nanoseconds it takes to get a few bytes from chacha. except that iirc ~~chacha is not seekable~~ (edit: chacha...

> Before worrying too much we should actually measure how many nanoseconds it takes to get a few bytes from chacha. for comparison against squares-rng, for the squares32 variant I...

I think this would be very useful for domain-specific languages embedded in Rust, e.g. in [fayalite](https://crates.io/crates/fayalite), I've implemented field access for `Expr` (a struct representing an expression of type `T`)...

a new standard library use-case [just came up](https://github.com/rust-lang/portable-simd/issues/459): converting `Simd` to `Simd`, also for enums.

The standard library just uses bytes everywhere else not because Rust knows that's the correct thing to do, but because that's the correct thing to do for all *currently implemented*...

> Perhaps a debug-assertion and/or eprintln that warns if we see repeated reads of under (say) 4kb on a File? you'd probably also want to check if it's a tty...

I'm assuming `impl BorrowMut for &Vec` is a mistake since that would allow creating multiple mutable references to the same slice at once in safe code.

finally a way to compare things in standard algorithms where you need extra state outside of the items you're comparing!