I am truly baffled.
I am a programmer who primarily programs in Rust. I first noticed this problem when I noticed Cargo (Rust's package manager and build system) was having trouble downloading dependencies, failing with a connection reset error. Curl exhibited this problem too, although my browser seemed to be unaffected. Examining the output of curl -v I learned that DNS resolution succeeded, the connection to port 443 was successfully established, and then Curl got an ECONNRESET before sending the first byte.
After doing more testing, I noticed that Curl was able to access google.com, stackoverflow.com, python.org, ruby-lang.org, rust.facepunch.com (the site for the videogame), and lib.rs (a site detailing Rust libraries that is not itself affiliated with the Rust project) with no issues, but was blocked trying to connect to crates.io, rust-lang.org, rustup.rs, or docs.rs. Stranger still, connections to these blocked sites would succeed seemingly at random approximately 17% of the time (measured by a batch script that tried over and over 1000 times in a row), and that once a connection had succeeded it could remain open indefinitely. After probing further still, I noticed that connections from Firefox and MSEdge would get reset too, but only when accessing the http version of the website, not https (curl and cargo were blocked on both). None of the other machines on my network are affected in the slightest. I have no explanation for this.