now :: random
google-scholar orcid linkedin

Is my goal in life is to be good at things?


This was written over the course of two non-consecutive days, as an initial outpouring of thoughts followed by a more stream of consciousness reaction.

11/01/26

Liking being good at things, is that why I like learning? Do I just want to feel accepted, in an in-group? I want to signal to others that I can do the programming, when I haven’t done it in a while, that I am a programmer who enjoys the artisanal vibe of it, when really I am just vibing through it because to me, the ideas are more important than the exact execution?

What would I do if I wanted to use an ESP32 microcontroller to do a simple task? How much AI would I use? I believe it would be a function of how much I cared about the output, right? The more I care about the output, and the faster I therefore want this thing to happen, the more likely I am to use AI, and the less likely I am to care that I’m not learning or retaining any of this.

Which is interesting, but I guess in a way it makes sense that my desire to truly sit down and learn something is inversely proportional to the relative urgency (or even just perceived urgency) of a given thing, simply because learning takes time and effort, and to learn deeply cannot really be cheated.

To “learn quickly”, to “get up to speed” with things, to context switch rapidly and have a surface level understanding of many things, this is a strength of AI, but I feel my mind slipping when I use it. I learn things and then the homework is supposed to reinforce these ideas, but if I forgo doing the homework myself, I have lost out on the art of learning. And yet, if that homework takes away from my ability or time to do research, then I am simply less likely to do it, or even feel bad that I have internalized less, since it matters less to me in the long run.

The problem in my mind is that I want to be a part of this in-group, to signal in some way that I’m still capable of programming with my brain alone, even though with every passing day I feel that less and less. Similarly, and perhaps more of note, I have struggled (relatively speaking) with math for a while. It’s not an intuitive subject to me, especially as the symbology and proofs become harder and more intense, and more shit flies up in front of my face and I go “huh? what the fuck is going on?”, and then shake my head while copying it down.

I’ve experienced some variation of that exact reaction in five math classes during my undergrad, none of which I received an A in, and the end result is that I had to cut corners in my understanding - despite my best efforts, maybe proofs are not for me? This trend has continued in grad school, where I lament loquaciously about the labors (of love?) in my randomized algorithms class. I am quite bad at following the proofs.

Something I tell myself is that if I had more time, I would be able to digest these proofs better than I do currently, but I’m not sure if that’s true. In a similar way, while I reflect on using AI to help me progress with these problems, I claim to myself, pearls clutched, that I could do it by myself if I had enough time. How much time is really enough? There are days when I spend hours working on my own, to no avail. There is something to be said about the bag of tricks mathematicians accrue, the hallmark of a “good” mathematician, or even an average one.

My bag is freaking empty bro.

Sometimes I think that’s true for my CS/programming knowledge too, but for this I am slightly more confident in my ability to recall things I’ve actually learned before, and to draw connections across disparate subfields. It’s a little easier also because inherent to the art of programming, for quite a while, was the need to write the code. In math, the crucial difference is that I can read a textbook without doing a single practice problem, and feel like I’ve gleaned the world, when in reality my mind is as porous as granite.

Too many shortcuts, “too little time”, and you have a recipe for disaster.

And I really want to be better at math, I really want to revive the joy I had with programming, but I realize that there are only particular parts that I care about, and then it all gets washed away. There’s a core of a problem, an essence, where if you automate the scaffolding, just maybe this itty bitty part will pique my interest and have me focused on it to a degree where I feel like I have the drive and enjoyment to figure it all out on my own. The real world, the world of real, useful, scalable software, is full of the scaffolding, and so then I end up abstracting more and more of the coding, and more of the design, and more, and more, and more, and the finished product resembles a mere fraction of my initial idea, haphazard as it was. What’s up with that, I say, wondering where my thoughts went.

11/14/26

This update is a bit more rambly, and touches not only on being good at things, but learning, an interlude on motivations and writing this, and time to learn.

The other thought is that perhaps I just haven’t found the best way to learn yet, and there’s no shame in that. Whether I believe myself at this moment in time - that there’s no shame in knowing how to best learn something - is up for debate, but regardless, I am in a personal renaissance of moderate proportions whether I want to be or not. It’s odd, reflecting on something I wrote two weeks ago and seeing both echoes of truth but also falsehoods and silly phrases being peddled. When I said my bag was “freaking empty, bro”, was that relative to, like, mathematicians?

Relative to high school me, I am killing it. Relative to college me, still killing it. I remember the time I first learned about doing structural induction nicely in my graph theory class; before that, I was getting punked by the type theory proofs for my research even after staring at TAPL for a while.

As an example for me realizing that I might not have found the best way my brain learns just yet, this past week, I stopped taking notes halfway through 271, Randomness and Computation. Usually, this feels like a no-go in my soul because what if I miss some important detail or explanation in a particularly nitty-gritty proof, or something similar? What I realized, slowly, was that the most important things to understand most of the time were (a) the proof strategies, i.e. how the proof worked in an overarching way, what the lemmas did, how they connected, etc., and (b), some very specific mathematical jumps, like why exactly we get to say the expected number of monochromatic $k$-cliques in $G \in \mathcal{G}_{n,~ p=\frac{1}{2}}$ given edges either black or white is $\binom{n}{k}\cdot2^{1-\binom{k}{2}}$.

A $k$-clique is a complete graph on its own, and has $k$ vertices and (crucially, from being a complete graph) $\binom{k}{2}$ edges. In the $\mathcal{G}_{n,~p}$ random graph model, each edge exists with probability $p$, so if $p=\frac{1}{2}$, then we first choose $k$ vertices from $n$ (the $\binom{n}{k}$), then we notice that for each edge, there’s a 50/50 shot it shows up, so that’s $(\frac{1}{2})^{\binom{k}{2}} = 2^{-\binom{k}{2}}$, and we have an extra factor of 2 at the front for each of the possible colors.

Stating that as fact is a bit rough, but deriving it as above helps internalize why that is true. Unfortunately, despite this being how my mind works, and the level of detail it requires to truly understand something, rarely does this happen in class. It’s an out of class exercise.

Crucially, in the before times, I was simply writing almost everything without discretion, which has sort of worked for me till now, but our prof is lovely and yet a speed demon with his words, so if I was focused on writing, I’d only really catch the tail end of the explanation of another important thing. After stopping my fervent note taking, I lost a lot of the mathematical rigor in my own notes, but I was able to follow the flow and jot down specific notes/derivations as needed, after processing them for a couple seconds. This does only work because I know that typed lecture notes are released after class, but that’s still notable.

Writing this out, it sounds a bit silly that I hadn’t tried it, but it’s harder than it seems to break habits, especially if they’re not inherently bad, and have worked in the past.


(an interlude, regarding motivations for writing and publishing it)

Will I share this publicly? I’m not really sure! It’s an interesting thought to realize that if I do, someone may read this article after finding my website, and realize that I struggle with things. I do think I’m fine with that, personally, but it’s not just about my acceptance of self (although perchance that is a nontrivial factor, my convalescence), rather this impacts my public image in a sense, my admission of weakness in the public eye.

On the flipside, I think this has the potential to be motivating. At this present moment (subject to change), I’m able to admit to myself that I am successful, at least a little bit. I’ve done something right, done good in high school and college, and in the purely work/research/education achievements way, I am doing a good job. In this way, I could potentially be a role model - I’ve already had people email me and ask for projects, advice, and so on. So to see someone who has succeeded openly say they are fallible could be meaningful or impactful in some way. Maybe even just by a little bit. As much as we compete in this rat race, scurrying up to the upper echelons, at times becoming increasingly jaded, we should also take a step back and relax and allow ourselves to fail with grace. I know I have to.

It is curious, given the image I ideally want to portray of myself is not of a firstie PhD student going through his (hopefully) quarter life crisis, but rather that of a reasonably competent person who talks about both his personal and technical interests, writes creatively and not, et cetera. And yet, I don’t think I’m good enough at the technical stuff? Or what I have to say is perhaps not worth it? I dunno! There are many things to think about w.r.t the “value” of writing something, the utility it brings, both to oneself and to those who read these things. I personally find a lot of utility in writing about my feelings, views on the world, ruminations on art, AI, and technology in that abstract way, and I separately derive a decent bit of joy from creative writing as well, when the ideas hit me upside the head with a skillet (/pos). As of yet, there have not been many instances where I truly derived joy or understanding (or both) by explaining technical concepts, especially concepts that already have fantastic explanations elsewhere. Perhaps someone will gain elation from my collation of resources, but that feels silly. Even if it may be valuable, it is less so to me, and it feels like I’ve done less to warrant even putting it up on my own website.

The clear exceptions to this general rule are already up on my blog: from Advent of Code 2024, Lean, and LLVM to name a few (is that all of them thus far?)

What technical stuff would I like to write about? Would writing it up benefit me more than I think?


post-interlude

Returning to the notion of being good at things, goals, and so on, I present the following quote:

A jack of all trades is a master of none, but oftentimes better than a master of one

– William Shakespeare

Ignoring the fact that William Shakespeare said that (I just learned that today), I believe this quote to be one of my personal mantras, which is a pretty interesting choice for my current career choice. At the end of a PhD, you’re supposed to be an expert in your subfield, the bona-fide go-to guy for something you worked on. Are these two ideas mutually exclusive?

Separately, I have also thought about this quote in a different way, butchering it to consider a person who is both a jack of all trades and a master of {one, some}. Does this exist? I distinctly recall my college admissions officers talking about “pointy spheres”, this view that the people who succeed (in the process of admissions or otherwise) are usually both good at most things at a baseline, and have some specific set of things that sets them apart from the rest of their peers. It’s certainly doable, I’ve seen some insanely cracked people around in my time, excelling at everything and being among the best in the world at a smaller subset of things as well - I wonder if it’s a realistic goal at all.

If the goal is to be good at things, by proxy, this is the goal… at least until Goodhart’s Law kicks in.

I think a part of the reason for this being my mantra is not just that I’d like to be good at things, but rather the rush that comes from being good at connecting ideas together. In a way, this is how I do research, finding unique ways to conjoin techniques to make things work better - and applying that idea generally is equally cool to me. As an example, the overlap between people who understand the LLVM ecosystem and quantum computing is small, and I think it’s really cool that I have the abilities and knowledge to be one of those people. Ironically, I got accepted to the lab of one of the only other people who does; he’s pretty cool, as are his students!

A separate question that I’ve been thinking about recently is this notion of classes helping me focus and learn better. I’ve always believed this to be true, especially up to and including high school, where I felt like it was a well-paced learning environment that made me practice intentionally enough that I internalized concepts, ideas, and techniques reasonably well regardless of discipline. Notably, however, high school was both (1) more structured than anything else that followed, in a good but non-negligible way, and (2) gave everyone more time to digest information, i.e. the full year per class, typically.

As a brief tangent, the more I think about college’s semesterly or trimesterly or quarterly time blocking, the more I feel that it’s occasionally wild that we still succeed. I recognize that my personal learning style of drilling down to the base levels to understand things from the bottom up sort of necessitates more time, and thus with strictly less time (semester vs a full year), it becomes strictly harder. I wonder how things would change, for example, if we had a full year of classes, and had more practical coursework but also more time for each practical assignment. Some of my favorite assignments in my later classes were in computer systems courses like my grad compilers course and advanced computer architecture course, where we were given a month to work on certain large scale projects and homeworks - I incidentally wrote about one as well, the other was a Tomasulo’s Algorithm Out of Order Processor simulator, written in C++. Being able to entrench oneself in these ways seems like a path to learning better. In a similar way, I would imagine (though I have no real idea) that I would do better with more time to solve math problems, perhaps. Broken record.

But this comes to the question of do classes actually help focus, or is it just more time? I think the accountability that classes gave me helped me focus, and nowhere was this more true than over this past summer, when I tried teaching myself some quantum error correction theory. There was a bunch of disparate math to understand, and even after finding some really good resources that were rather well designed in their structure, it was still intense. I got further than I thought, but I felt like I’d sped through the textbook without internalizing; in this race to get to the next topic, for fear of a lack of preparation prior to my PhD, I’d just continually set my sights ahead of where I was and chugged along. Classes come with assignments, lecture notes, an intelligible lecturer that’s also not going to allow me to ask for the answers (re: AI tutors), and a timeline that is reasonable.

All this to say that I think this is something to think about when designing the next generation of curricula for (higher) education. I know my own advisor participates in key discussions on this front, so I’m curious what he thinks about this. If the goal is to have competent scientists, to have good engineers, to have those that understand the fundamentals deeply, this seems like a referendum waiting to happen (not even necessarily in the way I suggested).

In the conversation that was the impetus for this concise treatise (say that five times fast), I asked myself, “Since being good at things requires learning, do I even like learning? Or is learning simply an obligatory trial to trudge through on the path to ‘good’ness?” Now, past me forgot that the word “competence” was a thing, but he was going through a bit of a crisis so we’ll be polite and let it slide. There are a few things that, for better or worse, I have bound myself to – they identify me. One of these is my adaptability, a la jack-of-all-trades, my ability to learn quickly and get good at things. The second was that I believed I really enjoyed learning, and that this was one of my superpowers. It is in some ways conceited, I acquiesce, because that surely isn’t unique to any one person, and yet I felt that in this sea of people, I was able to do these things better than the rest. If these aren’t true, who am I?

The logical side of me knows that as long as my overall outward attitude towards learning is the same, then the motivation may not matter at all. The results will keep coming, surely, and I’ll still succeed and get better, which is (for sake of argument) exactly what brings me joy… but that feels wrong?

I overthink things quite a bit, and I’m not gonna stop now; I wonder what it means to even like learning in the first place. Like my brother would say, I’m sure “there are levels to ts”. I can at least say that I hate feeling sedentary and I’m always learning something. I chose a field that requires it, and my approach to many things in my life is to get better at them, at times (most of the time) simply for the sake of it. In “Doing things because they are hard”, I think I touched briefly on a good, salient point, that doing hard things can be good simply because the journey is longer, and it affords us the opportunity to learn more, glean more, gain more, regardless of if we reach the end destination or not.

I have yet to finish my QEC textbook, and I’ve reread several sections thrice, even four times over to internalize them. The journey is tough and the end destination isn’t in sight, but through that journey I’ve been able to delve into some awesome research. That slog was and still is an opportunity.

The other point I made in that blog post, around a year and a half ago, was that doing the hard thing and completing it separates yourself from the riffraff (I didn’t phrase it as callously last time, but riffraff is just a really good word. Watch the movie “Pixels” for a good use of it). Inherent in this is the idea that in elevating myself above the rest, I am basing a life goal and with it, perhaps some nonzero % of my worth on a metric relative to other people.

Comparison is the thief of joy

That’s like the $n$th time this has been on this blog, but it’s always helpful to have a reminder. Every time I go back and read some of my old writing I always think I’m smart, it’s funny. Repetition helps, I suppose. A natural extension of this idea is the initial phrase, “my goal in life is to be good at things”. To be good at something implies that another person is bad at it.

Ironically, this reminds me of the probabilistic method - the randomized algorithm to find a MaxCut in a graph returns a cut with expected size $E[X] = \frac{|E|}{2}$. In order to show that $E[X] > \frac{|E|}{2}$ strictly, it suffices to show that one possible random assignment of the vertices groups all vertices in one vertex set, and none in the other. Thus, the size of the cut is 0, and in order to balance this out and maintain $E[X] = \frac{|E|}{2}$, there must be a cut whose size is strictly larger. $\square$ For more fun with this, check out my formally verified version!

Regardless, these are relative metrics, and the goal should probably be updated to involve as little relativity as possible (sorry Einstein). There are many deadlines I have towards my learning, most of them being somewhat arbitrary owing to the odd nature of my work and degree in comparison to a more traditional career path, and so with this arbitrariness comes the freedom to relax my self-imposed, partly imagined deadlines. Sure, that conference deadline isn’t changing any time soon, but whether I finish my paper is separate from when I feel like I’ve sated my learning in a topic to the point where I am confident in my understanding of it.

Right now, I’m trying to teach myself the Lisp programming language. I found a textbook that seems nice, if a little intense, and I’m going through it! I’ll try picking a small project soon to see how much I understand, and hopefully some progress shall be made. At worst, there is a slight deadline on when I learn this language (I should ideally know about it for the programming systems prelim), but that’s next October, so the new goal is to practice learning without artificial deadlines.

I think an even better example, one with truly no deadlines at all, is my desire to learn about art. I’ve been reading and consuming media about it at a slow and steady pace, and I think it might be a good barometer for checking in on myself in a few months.

I’m curious if anyone else reading this (assuming I put this out) has felt the same way. Email me, I’d love to talk!

Thanks for reading!

– Arjun