By the time I’m starting this blog, it’s August 28th and we’re wrapping up GopherCon talks. It’s been a long week of talks, social events, and me frantically preparing for a talk because someone thought it was a good idea to accept me (a Gopher with barely even 1 year of experience) and to let me on stage in front of hundreds of more experienced engineers. In all honesty, I was not expecting to be accepted to talk, but I was, and now here I am.

My talk, “Invisible Insight: Strategies for Auto-Instrumenting Go Applications Without Code Changes”, was a journey that stretched over nearly 2 months. I went from someone who had never touched auto-instrumentation to someone who kind of, almost, maybe understands the ins and outs of compile time and run time instrumentation in Go. I would not claim to yet be an expert in this area, because I think it would take many more months or maybe years to reach that kind of level, but I’m 2 months smarter than I was before, and that’s all I’m going to ask for.

Auto-instrumentation is, as it turns out, and entire can of worms that I did not realize I was getting into. For starters, I can’t even use most other programming languages as as template for what happens in Go. Since Go is a compile-time language, other approaches to auto-instrumentation like eBPF, library injection, and annotations are difficult and sometimes impossible. The option that we, at DataDog, opted for was to use this flag called -toolexec. It’s a funny thing that, in very simple terms, allows you to wrap the standard Go compiler (and other tools) to do extra, self-defined work for you. And when you bundle up that self-defined work into a library that you can run with a simple script, you get our tool, Orchestrion.

My talk focused mostly on Orchestrion and how it compares to the other big auto-instrumentation approach we have in the Go community: OpenTelemetry’s eBPF (extended Berkeley package filter) library. As I alluded to earlier, eBPF is a runtime approach that I, at first, expected to be impossible with Go. But turns out, when you’re determined (and smart) enough, eBPF does work. The only limitation we have with Go, once you boil it down far enough, is that we cannot make changes to our source code at runtime. So what eBPF does is to make changes to the kernel at runtime. Really mindblowing, if you ask me.

So then, what’s the difference? You can watch my talk to find out, or I can just tell you now. There are a couple of main differences:

Ta-da! All the conclusions I came to during my talk have been boiled down into that table.

More than just eBPF, toolchain and compiler systems, and telemetry knowledge, I’ve come out of this process with a whole lot of presentation learnings. I’ve done a fair share of presentations before, but they’ve always been small scale. Think science fairs, mini workshops in college, etc. Never once have I spoke to hundreds of people at once, live. Much less with a live demo that was not supposed to be live until about 10 minutes before I got up on stage. Because presentations never Go as you expect, and this one was no different.

Turns out, cramming as much as I can about an entire field of instrumentation is hard. There were so many details that I missed (how could I forget that a Mac does NOT run on Linux???). I’m incredibly lucky to have my amazing coworker, Kemal, who helped me throughout this entire process and helped sanity check my results. He even donated his machine to run my code on a Linux VM because my machine would not behave. It was his demo that got shown during my talk, so huge shoutout to him, the secret supporter behind the scenes.

And of course, as things always do, things went wrong. The demo didn’t work. Or at least, it was working, but it was not giving the values that I was expecting to see. Numbers were a little off, and I have no idea how navigate it. For a brief moment (that felt like around 30 years, but was realistically around 3 seconds), I could only freeze and go through every single option I had. Bluff it off? I’m not nearly tenured or experienced enough to pull that off. Admit things were wrong and move on? Yeah.

After I got off stage, I could only sigh and immediately text all of my friends with the most dramatic retelling of how everything fell apart on stage. Because I love to exaggerate and to tell stories. But it was fine; it was a risky move to try showing things live, and it was a risk I wanted to take. It didn’t work out exactly like I wanted it to, but that’s fine! I can only accept that and move on in the moment.

Is there more to do? Absolutely. There is still more in the demo that I want to improve, and I want to get my Linux VM working on my machine so that I can actually see the results directly on my screen. I could make the demo more realisitic, rather than the small, black box example I currently have. I’m planning to work with Kemal to build this up. Shoutout to him again for teaching me all I know about instrumentation, and hopefully we can present the updated version of this talk in another conference!

Aside from my talk, GopherCon was a lot of fun! I’ve met so many Gophers from across the world and re-met several I’ve known. It was great seeing Katie, my former manager at Datadog, giving her talk right after me and Felix, my coworker at Datadog, going on stage right before me! And it was fantastic helping out at the Asian Alliance meetup, even if it was just to order a whole lot of sushi for the group.

At time of writing, the recording of my talk is not yet up on the Gopher Academy YouTube channel yet, but it should be uploaded sometime soon. But if you’re reading this and you’re interested to see the code and to learn more, my GitHub project, including all of the code and my slides, is here.

Am I proud? Hell yeah. It was a really long process, and I didn’t write that much in the eleven months that passed between one NaNoWriMo to the next. That meant that I felt a little out of my element going in, but at least I had the foresight to at least open the doc before the first day so that I could reorient myself. What was I writing again? What chapter are we on? What’s supposed to happen next?

Writer’s block is weird that way. I ended NaNoWriMo ‘23 at a weak point. I was in the middle of a transition, with no clear scene for me to work off of. It was a whole lot of character development that I needed to write but could not decide how to. There were scenes that needed to get done that I couldn’t get to, simply because I couldn’t figure out how to connect one scene to the next without feeling as though I was dragging myself through mud. So when I finished the month, it would be months before I could bring myself to open the doc again, and many months before I could write a few paragraphs, much less an entire scene.

So, I knew I had a challenge this year. And I’m a perfectionist. I hate leaving things undone, and I’ll painstakingly agonize over every word as I write until I’m at least somewhat satisfied with how the sentences flow and what the characters are saying. Does it fit their personality? Would they actually do this?

And that was the problem. Let’s be serious. I’m writing the first draft of a novel, and it’s the first time I’m writing something of this length in… ever. I have never written anything this long before. Do I logically think I could write a Pulitzer prize winner on the first draft? No, of course not. Does that stop me from trying? Also no.

It’s two very different things to recognize this fact about myself and to actually do something to fix it. I spent the few months leading up to this November anxiously trying to get myself into the headspace of a “first draft writer.” We’re allowed to skip entire scenes, leave comments to remind ourselves of what we need to work on later, and keep barreling on so we don’t lose steam. The entire point of a first draft is to make it exist, not to ship it out to publishers.

Realistically, it will take me another year to finish the first draft if I keep the pace I’ve been going at. And it will take me years to edit and to write a second draft. And even more years to write a third draft. Rinse and repeat. And from there, if I decide I want to publish, it will take me many more years.

That’s reality. And recognizing this made NaNoWriMo so much easier.

So I wrote a lot of “”’s this year. And left a lot of “BLANK”s when I didn’t want to stop and come up with a name for a new character or city. And when I say a lot, I mean a lot. There’s entire chunks of time (in story) that are missing, where I have mentally filled in the gaps of how the characters have developed in that time. And that’s okay. I will come back to it later, once I am done with this draft.

It’s frustrating to admit, but yeah, I made a lot of progress this year by not writing.

But isn’t that so satisfying?

I could write a whole novel (as scary as that is) about how I feel about NaNoWriMo as a program. I love it as a way to encourage myself to write, and their pretty graphs and awards do wonders to give me unimaginable amounts of motivation. But unfortunately, in the final few days of NaNoWriMo this year, I learned that the team is what they call “AI neutral.”

There’s a time and a place for AI. Like in cancer research. But in all honesty, I do not believe there are many other situations where AI is genuinely useful. Generative artwork? Absolutely not. Chat bots? Maybe I’m old fashioned, but I like Google just fine. And it’s not just that I think there are better alternatives. All things considering, AI is fantastically helpful sometimes, and I admit I do use it when I absolutely cannot find what I need on Google.

What I hate, though, is the “hidden” impact AI has on us. Stealing data from people, especially artists and creatives, to train their models without permission. Draining the environment of resources to power servers that run the models. The humans behind training data who are underpaid and not recognized. These things weigh heavily on me when I think of AI and whenever I see a piece of AI generated art.

So when I heard that NaNoWriMo is “AI neutral”, I could not help but finish this month with a sour taste in my mouth. As a creative community, NaNoWriMo was the last place I expected to be welcoming of AI works. It stings a little, of course, to have spent dozens of hours writing scene after scene, training myself to push past my too high standards, just to find that an AI generated work could do all that in a few hours, or maybe a day, with barely any human intervention.

What’s the point of that? “Writing” without emotion or human thought? What message could an AI come up with that isn’t an idea stolen from other writers? Was it Shakespeare that it trained off of, or maybe Agatha Christie? Or was it me? How many gallons of water did it require to generate 50k words? How much of that water are we getting back?

Anyways. Aside from doom spiraling about AI, I spent time thinking of what I wanted to do next year. I don’t think that I can, with a clear conscious, go back to using NaNoWriMo’s site. I have already found a few alternatives, including TrackBear, which I’ve been using a little since November ended. And, of course, I’m a software engineer, so I have the means to create what I want to keep from NaNoWriMo myself.

So, when in November 2025, I am no longer a user on the NaNoWriMo site, know that this is the reason why.

Lately, I have been obsessed with the idea of trying my hand at book-binding, and for once, I might have some minimal qualifications. They are as follows:

1. I read a lot as a kid. Specifically, I read many children’s books from our local library that were not well taken care of, which meant that I could peer down the spine and examine the binding, pick at the covers, and spread the pages far enough until I could see the stitching. So if you really think about it, I’ve known the general anatomy of a book since I was around four years old.
2. I’m currently taking 21H.343, or “Making Books: The Renaissance and Today”. That means that I’ve tried my hand at printing using a printing press and examining old books and incunabula. I’ve learned the basics of how books are made and what components they are made of, which will definitely help in this journey.
3. I said so.

So, with all that I said, here is a recounting of my book-binding journey.

I decided to bind a final project that I did senior fall for the class 24.917 (Conlangs: How to Construct a Language). As the name of the class suggests, I had to create a language from scratch for my project, including its phonetics, grammar, and syntax. It was one of my favorite classes at MIT, partially because of how much freedom there was and partially because of how much I learned, and I loved it so much that I decided to use the language I created in my NaNoWriMo ‘23 novel. And since this novel (and thus the language as well) is still very much under construction, this 27-page paper sounds like a great option to print.

In terms of page size, there are a couple of options. The most common ones are folio (f°), quarto (4°) and octavo (8°). Basically, folio means that we fold a piece of paper once before printing, so each paper will give us four pages overall. Quarto means we fold twice, so we get 8 pages. Finally, octavo means we fold three times, so we get 16 pages. Of course, the more we fold, the smaller the page sizes will be. I wanted the book to be small (and to save paper), but not too small, so I decided to go with the quarto format. Each piece of paper will give me 8 pages. Unfortunately, 27 is most definitely not a multiple of 8, so I’ll have some extra pages at the end that I’ll either have to cut off or use as blanks.

Maybe the most annoying part of this process was figuring out how to get these pages to be print in the correct location and orientation on the page. As the image from Wikipedia shows, getting the pages in the correct order for when we start folding and cutting means printing pages 1-8 on different sides AND printing some upside down. There are some options online for organizing your pages in preparation for printing, but what’s the point of taking on a project like this if I’m not going to do it all by myself? Besides, I’m a computer science major, so something like this should be easy to code up (refer to the first sentence of this post).

There were a few things I wanted my code to include: choosing a format (folio, quarto, or octavo), how many leaves per gathering (AKA, how many pages should be folded on top of each other before sewing them together), and if I would like extra pages at the front and back. The reasoning for the extra pages is because (at this point in the planning process), I think I could use the extra pages to attach the inner gatherings (pages) to the cover of the book. The existing site I linked above has many, many more settings that I may attempt to recreate in a future iteration of this. But given how small my binding project is, I really didn’t need all that.

Anyways, this is how the site turned out:

I did not put much (if any) CSS into this <3 (for now). The site is simple; you just upload a file, select the format you want to print it, whether if you want extra pages (pamphlet) or not (book), and how many sheets we want per gathering. As I mentioned before, I decided to go with a quarto book. And since this is such a small project, I decided to keep only one sheet per gathering.

The result of clicking the download button was four sheets of paper, each with 8 pages on it, which I could then fold into the gatherings I needed:

Now, to combine these pages together, I decided to use a coptic stitch (because it was the only stitch I learned to do in my 21H.343 class). To do this, I also made rudimentary front/back covers out of some plain white cardboard, which I cut to be just slightly larger than the individual pages.

Coptic binding takes one gathering at a time and ties them together in a stitch that looks kind of like a braid. It works by piercing holes through the spine (here, the folds in the paper), passing through the spine of the next gathering, then tying them together. It’s easier to use a curved needle and waxed thread to do this, since the needle allows you to get the angle you need to work under previously set knots and the wax keeps the thread closely attached to the paper. Due to a lack of resources, I only had a simple sewing needle and thread to my name, but it wasn’t too much harder to do this.

I forgot to take a picture of the final binding before I made a mistake that utterly ruined this project (explanation coming), but here’s a picture from the Wikipedia page:

As you can see, the stitches and the binding are visible, which is one of the aesthetic benefits of coptic binding. Unfortunately, I did not really want the bindings to be visible; your typical book doesn’t use a coptic stitch, and thus bindings are hidden behind cloth and wax. I decided to hide the binding by simply cutting a piece of cardstock and gluing to the spine.

At this point, I made a critical mistake. In my attempt to save paper, I tried to get this piece of cardstock from the scraps that I had left over from creating the covers of the book. But there were two main issues: all remaining scraps were too short and/or too narrow. I didn’t have enough surface area to glue the paper down to the spine without creating a ginormous mess on my desk.

After a brutal struggle, I managed to figure something out by taking the widest piece I had, gluing quickly, and then stacking heavy items on top as I waited for the glue to dry. This managed to cover up the evidence of my past attempts (which resulted in ripped or fraying paper) while giving it (almost) the effect I wanted. Here’s the final result:

(You can still see the dried glue from previous attempts on the spine, but we can ignore that)

This attempt was far from perfect, but it was an interesting project that really pushed my creative skills. Overall, I kind of like how it turned out! It’s a solid first try, and a good learning opportunity for future attempts 👍

Regardless, I wanted to spend my first blog post talking about NaNoWriMo! NaNoWriMo stands for National Novel Writing Month, and it happens every November. Writers, both recreational and professional, are encouraged to work on their novels for the entirety of November and write 50k words. It doesn’t really matter if you start a project from scratch or continue an existing one, but for me, I decided to start from the beginning.

I had a novel idea for the past little while, but I purposefully procrastinated on starting it until NaNoWriMo started so that I could fully encourage (re: force) myself to work on it. The website has a great portal for entering how many words you’ve written in a day (as well as how long it took, how you felt, etc, but I didn’t really keep track of that, since I tended to write throughout the entire day). I also made an effort to post on my Twitter about how much I was writing, so that I could keep myself accountable, but I also… stopped doing that.

I have always enjoyed writing since I was a kid – a lot of my earliest memories were of me writing my own picture books. Even when I hadn’t learned to write yet, I was narrating to my dad what I wanted him to write, and I would fill the empty spaces with illustrations. Back then, before I had ready access to a computer, I would write on paper, fold it up, and staple the spine so that I could open it up like a book. (This reminds me that I would love to try my hand at book binding!) But while I spent a lot of my early years writing, it kind of faded away as I got more busy in late high school and into college.

I’m super proud to say that I actually wrote an entire trilogy when I was in middle-high school. Was it any good? No. Absolutely not. There were about 10 different plots going on at once since I couldn’t decide which I wanted to keep. But regardless, I somehow stumbled my way to the end of a ~150k word journey. And that was one of the last writing projects I had.

But! I’ve always wanted to participate in NaNoWriMo, and what better time than in my senior year of college? I was lucky enough to have a job offer at DataDog lined up post graduation, and to be mostly done with my academic requirements, so I had more time on my hands for the first time in a few years, and, most importantly, I had inspiration.

The hardest part about writing is truly getting the story in your head into a fully fleshed out, developed idea. I’ve had many, many story ideas over the years, but they were such one-dimensional ideas – more like a single quote or scene that I wanted to write into reality – that they could barely be considered a wisp of smoke in the grand scheme of things. In comparison, I had slowly been brewing an idea that was more akin to a flame, or even an entire bonfire, in terms of my passion and my motivation for it.

I have a lot to thank to Madeline Miller, the wonderful author of the book The Song of Achilles. It’s a retelling of The Iliad by Homer*,* the story of Achilles and the war against Troy, from the viewpoint of Patroclus, who was Achilles’ close friend (and lover?). I read The Song of Achilles earlier in 2023, and I was absolutely in love with the depth of the characters and the inescapable tragedy in the end. Her writing is so captivating as well, and I’m currently reading another one of her books, Circe.

I decided that I wanted to write something akin to The Song of Achilles, in that it is also a retelling of The Iliad from Patroclus’ perspective, but that’s about where the similarities end. I made Achilles and Patroclus women, because I figured that women would be more enjoyable for me to write as a woman, and I created an entire nation(s) to set the scene in. I created my own gods and a new war to take place. The prophecy is mostly the same, though. The princess who takes Achilles’ place in my novel is doomed to die while fighting her nation’s war.

What’s more, I also decided to pull from more Greek mythology. My story is about women (or at least, it is in it’s rough state right now), and what woman in Greek mythology could I easily take inspiration for? I took an alternate interpretation of the story of Medusa. Many people know Medusa as the evil, snake-haired woman who turns everyone she looks to into stone. She’s cruel, unforgiving, and a monster above all. But, what I wanted to take was her origin story.

Most stories say that she lay with Poseidon in Athena’s temple, and was thus punished by Athena to have her hair turned into snakes and to have her gaze turn others into stone. But other interpretations tell a slightly different story. Medusa was a mortal, and she was raped by Poseidon at the temple. Instead of punishing her, Athena gave Medusa her hair and her eyes as a way of protection, to ward off any other men who may try to take advantage of her.

In the end, it’s still a tragedy, for Medusa is still, by most people’s definition, a monster. And she was killed for it. And that is the story of Medusa that I wanted to incorporate into my story.

Currently, it’s been about four months since I started writing my novel, and many more months since I started developing its ideas in my head. It’s still a far way away from finishing, but it’s the most excited I’ve been over a writing project in many, many years. I still get bursts of inspiration and plot ideas when I’m going about my day, and I have a whole list of thoughts that I’m itching to put into writing. I feel that it’s safe to say that my motivation to write has returned after a five year hiatus. Here’s to, I hope, even more years of writing!