© 2026 Reid D. McKenzie, all rights reserved.
So last time I just brain dumped a bunch of notes on the subject of the AI Memo 514. This time I’ll attempt to offer an analysis of the paper, and compare it to the Knight thesis.
This summer I’ve been an intern at Factual, and this is an experience report from the semiannual internal hackathon where Alan ‘amalloy’ Malloy and I experimented with using Alexander Yakushev’s Skummet fork of Clojure to emit lean(er) bytecode.
So ztellman posted some of his library and I was bored (and also more than somewhat curious what Zach reads) so I figured I’d make a list in the hope that at least I found some interesting reading material. It’s not complete but it’s pretty interesting as-is. I’ll happily take changes if anyone else cares to read the original.
This was originally a term paper of mine, written in markdown, and appears here because reasons.
In working on my Batbridge simulators and other adventures in computing hardware, one of the problems I have repeatedly encountered is the need to generate bit encodings of symbolic instructions. When working with “well known” architectures with existing tooling support, one can usually find an “off the shelf” assembler either in the GCC collection or elsewhere. However when you’re working against a custom instruction set you’re on your own for tooling.
Here is some of my motivation for Oxlang. I’m sorry it isn’t shorter, I gave up on copy editing to turn in. This was originally a personal email and appears almost unaltered.
When I last wrote about Oxcart work pretty much went on hiatus due to my return to school. As there has been some recent interest in the status of Lean Clojure overall I thought I’d take the opportunity to review the state of Oxcart and the plan for Oxcart going forwards.
This is also addressed to @lojikil towards whom I have threatened to write this post on multiple occasions. It is expected to piss off loper-os. Above all, it represents my opinions backed up only by a cursory research effort. You have been warned.
Recently, a friend of mine at ARM has been kicking around the idea that syscalls are inefficient and that a more than reasonable speedup could be achieved across the board in *nix operating systems by providing a mechanism for reducing the overhead of syscalls.
