Cameron

For the past two or so weeks I've been spending a lot of time writing Rust. Rust is a really spectacularly designed language with perhaps the greatest package manager.It's called Cargo. I've been continuously blown away by how lovely it is to write stuff in! Rust is fascinating to me because it's one of my first true compiled languages. In contrast to the other languages I'm somewhat good at: • R (mostly interpereted) • Python (interpereted) • C# (JIT compiled) • Julia (JIT compiled) I am only passingly familiar with C and have mostly repressed C++, so Rust was both new and refreshing to me. The workflows are strikingly different -- with Julia or R or what have you, I go through a very rapid iterative process. Write something, run it, see what worked, fix it, move on to the next issue. It's very slapdash. With Rust (and, I presume other compiled languages), I have to stop and think a lot more about what I'm doing and how I'm doing it. This is a bit more important in Rust because of the concept of ownership, an entirely bizarre concept of who owns what thing in any given program. My (admittedly frail) understanding of ownership is that everything that exists in a program is owned by a varible. Once that variable goes out of scope, the memory allocated goes with it. It's how Rust manages memory safety. It has a pretty steep learning curve initially, but once you've learned it it's almost effortless to use. All in all, it's a spectacular language. Fast, too. And ultra strongly typed with none of the rigid puritanical stuff that comes with Haskell.

# {{title}}, {{ pretty_date date }} I've recently been perusing an introductory text of information theory. Information theory is one of those things that I have always wanted to look into but never gotten around to -- but now I am an adult with real things to try not to do, so I figured what the hell. Information theory is the scientific study of how information can be quantified, stored, and communicated efficiently. It's fascinating to read the version of the book I have as the author (and Claude Shannon, the father of information theory) all viewed everything as telephonic or telegraphic. Pierce mentions computers a handful of times, but often only to describe what they are incapable of doing[^1]. Much of the content regards speech and how we can predict, transmit, and describe it. It appears to be primarily based on probability theory, and for good reason -- if you know that E appears 13% of the time and Z appears almost never, you can make asusmptions about how you code letters by assigning easily transmitted values for common symbols and more complex values for infrequent symbols. The primary reason I'm reading it is for the applications to financial markets. A lot of the reason why information theory came about was because there was a need for sophisticated signal processing techniques during World War II. Signal processing is something commonly applied to finance. Trades indicate resources, desire, risk tolerance, what have you -- but there's also a lot of noise. How do you tell people who are informed from people who are uninformed? How do you know whether a trade is actually meaningful to the long-term price of the stock? Dunno. Thought I'd read about it though. ## References *An Introduction to Information Theory* by John R. Pierce. [^1]: He calls them "automata", which is such a lovely old-timey phrase.

Forms of power and control and AI might have over us.

A look at Zed's new agent development tool.

Some notes about learning what the inside of my head feels like.