May I recommend... understanding Emacs's patterns
This is my sideshow in Sacha Chua's May Emacs Carnival. Enjoy!
This is my sideshow in Sacha Chua's May Emacs Carnival. Enjoy!
This has been requested a few times since I started making some demos and blog posts about completion in Emacs.
This page includes links to my Emacs config (zetta) as well as my dotfiles.
"My Emacs configuration, packaged as a small distribution. Around 320 packages wired up via a module DSL, with an Elpaca lockfile pinning every package to an exact commit, and byte- and native-compilation done up front so the first launch is clean.The name is a cheeky play on how we name certain minimalist text editors after Metric Prefixes nano (10^-9) or pico (10^-12). This maximalist editor config is named after zetta (10^21)."
"This post is the practical complement to all the other posts. Here, we showcase over a dozen workflows I use every day. Most are powered entirely by features that ship in the box with the VOMPECCC (Vertico, Orderless, Marginalia, Prescient, Embark, Consult, Cape, Corfu) packages, and there are 'Bonuses' which demonstrate workflows enable by 3rd party packages that build on top of VOMPECCC. The prose is deliberately thin, and you will find most of the demonstration is in the video below."
First off, this is all love. If you chose not to use LLMs, I 100% support you and I think that's cool. What I'm venting about in this post is the general unnecessary meanness I see in the threads of this subreddit w/r/t AI use.
There's so much dunking happening on posts that even mention AI/LLM/agent. I'm writing about this soon, but I wanted to dump some notes here.... I love 💜 questions about AI in this subreddit and I think it's a shame that people are getting admonished from this subreddit, which I hold very dear to my heart as I think it's one of the safest spaces on the internet.
I don't understand what's so transgressive about using AI or LLMs.
They are just tools. They are not some molecularly different entity or power from what we all use everyday (yes, even when we aren't programming). It is just a computer program. It is very effective, yes. But it is just, I repeat just a tool.
For me to believe something (like 'using AI is transgressive') it needs to be internally logically consistent. What must follow from a condemnation of LLMs is something like this:
"""
We're all typing these messages out on keyboards right? You're getting a lot of automation there. You barely need to move your fingers, spell-check happens automatically, you can re-arrange the text with editing features, and you can invoke a thesaurus at anytime. When you hit send, you don't have to physically transmit anything. Your computer does that for you!!! You're barely doing any work to communicate with someone on the other side of the globe.
Would it not be more virtuous to use a typewriter and send the correspondences physically, where we don't have these crutches? But still, typewriters offer a lot of automation. Symbols are standardized and perfectly rendered, they format text perfectly, with even spacing between letters and lines. So that must be transgressive too!
So wouldn't it be more virtuous to use a pen and paper? There you have to craft out the forms of each word's letter and ensure readable penmanship, which is extremely demanding of dexterity and fine-grain motor action. But pen and paper are also tools, and offer a lot of features. Paper is optimized to have pens and pencils glide along its surface, and pens are optimized for ergonomics and good penmanship. Pen and paper then, too, is surely transgressive.
Wouldn't it be more virtuous to carve words into heavy stone slabs? That requires a lot of time, craftsmanship, etc.... and you don't have the facility of a postal system to post your weighs-nothing paper notes for almost no cost. I could go on about the sins of how we write!
That was all about writing, but what about reading? A lot of folks don't realize that search engines have been using AI under the hood to enhance search results for years before LLMs came along (I'm looking at you Google's knowledge graph). Isn't it transgressive to use a sophisticated search engine where the content you're searching for bubbles to the top, and where you can write meaningful queries against the internet's full corpus of data? And that all that data has been collected, analyzed, and hashed for you by the SE provider? Sinful, I declare! Sinful!
Wouldn't it be more virtuous to go to a library and read the physical books for the content you need? But still then, libraries implement a sophisticated duo-decimal system, and are run by folks with Masters degrees in maintaining, organizing, and making accessible of the libraries physical data. So libraries, therefore, are transgressive too.
Wouldn't it be more virtuous to acquire your own library of books directly from the publishers? Then you would have to do the searching yourself, and then implement your own organizational system for making all your library's data as accessible as possible. Books in some ways are also a crutch. They are stable, well organized summaries(!) of the author's thoughts, and a lot of the work in discovery and comprehension has been done for you by the author. Also sinful, right?
So wouldn't it be more virtuous to correspond directly with the author (via stone slabs of course)? Then you need to bring an understanding to the conversation, formulate questions, and organize the information into a cohesive narrative. But, alas! You are simply being explained these complex ideas by the authors who have done all the critical thinking for you? Sin!
So wouldn't it be more virtuous to divine all this information yourself? Here we get closer to God....
"""
The point I'm making is that none of the above arguments are practical. There are fortuitous reasons why society uses these tools. Everyone's presence on this subreddit necessitates that they are using keyboards / the internet etc.... Then why are these okay, but not other tools?
Just started a 3rd read through after ~10 years since my 2nd.
This 👏 book 👏 is 👏 trashed 👏.
Front and back covers completely severed; stains from god know how many shitty university coffees; and lots of chicken-scratch hand writing, a good portion of which I can't even decipher anymore. Also many rogue question marks 😅.
Definitely well loved. I did think about giving this to the local library, but it's likely a biohazard at this point. Curious if anyone else has a more egregiously blown out copy than mine!
"This is the fourth post in a series on Emacs completion. The first argued that Incremental Completing Read (ICR) is a structural property of an interface rather than a convenience feature. The second broke the Emacs substrate into eight packages (collectively VOMPECCC) each solving one of the six orthogonal concerns of a complete completion system. The third walked through spot, a ~1,100-line Spotify client built as a little shim on top of those packages.
This post is the hands-on complement to the spot post. Where the spot case study reviewed a finished codebase from the outside, this one builds a tiny produce picker tool from scratch, one VOMPECCC package at a time. The use case is deliberately trivial: we have a list of produce items (twenty fruits and ten vegetables) with some metadata, and we want to pick one and do something with it."
"This essay's novel contribution to the critical literature is a typographic close-reading of one moment in Orin's morning chapter, where Wallace describes a peculiar feature of a Subject's handwritten note: "every single circle – o's, d's, p's, the #s 6 and 8 – is darkened in" (pg. 43). The argument is that the three darkened letters (O, D, P) spell, in Orin's perception, the name Oedipus. This may seem like a reach, but the encoding becomes the smoking gun in the case against Avril Incandenza when you appreciate Wallace's intellectual debt to Douglas Hofstadter and Gödel, Escher, Bach – a debt the essay documents in detail below."