For a long time I was, by my own honest accounting, a very well-compensated specialist with a B-minus in everything else — couldn't sing, couldn't really work a room, wasn't especially charismatic. So I did the only thing I know how to do: I treated the expressive half of myself as a project. The strange part is that it turned out to be one project, and the thing holding all of it together was learning to like my own company.
The B-minus, and the self-class that fixes it
The diagnosis came in New York, at the exact moment my outside life looked its best. I had the salary and the city and almost no idea how to be a person in a room. The way I put it to myself was unflattering and precise:
“I could not sing. I could not dress well. I was not particularly fit… I was not especially charismatic… I was a very well-compensated specialist with a B- in everything else.”— how I became superhuman in 2 years
So I decided to treat the next phase of life as a full-time attempt to fix every single thing I was bad at — and singing, improv, charisma, hosting, sound design all landed in the same list as the fix, not separate bucket-lists. The label I gave the result is the one written at the top of the D&D character sheet I keep for myself, and I mean it more than I'd admit at a party:
“Class: The Modern Bard — artistic with a touch of technical excellence.”— my own character sheet
The bard is the performer-generalist: sings a little, talks a lot, charms the room, tells the story. That's the whole self-class, and most days it's the part of me I like best. It isn't a list of hobbies stapled together — it's a single deliberate answer to a single thing I noticed about myself. Where the Maker is my engine pointed at problems out in the world, the Bard is that same engine turned around and aimed back at me. That sounds tidy and ambitious. It was also downstream of something much less tidy, which I'll get to, because it's the actual engine and pretending otherwise would make this page a brochure.
Singing, from level minus-999
I started genuinely tone-deaf — not modestly, actually could-not-find-a-pitch tone-deaf. I made a public video about thirty days of it, titled, with full self-awareness, "I learned to SING for 30 days starting from level -999." I'll cop to the clickbait before you catch it: I didn't practice every day for thirty days; it was inconsistent, and the title knows it isn't quite true. Wins and flukes, both on the table.
What's real is that I took actual lessons.
“I took singing lessons starting from what my teacher diplomatically described as a very low baseline.”— how I became superhuman in 2 years
The on-camera version of me is dry about the whole thing — when the day-five script gets to the part where I'm supposed to demonstrate, I just refuse: what, are you waiting for a demonstration? I said I've been doing this for a few days, I'm not gonna get it right. You want me to try anyway so you can laugh at me? Ok fine. That's about the truest portrait of the project there is: visibly bad, doing it on camera anyway, narrating the embarrassment instead of hiding it.
The home note — my one real singing idea
If I've contributed one honest idea to the how-to-sing pile, it's this. When you can't reliably hit any pitch, the move isn't to attack a whole song — it's to build one note you can find from anywhere, then learn everything else relative to it. I practiced a single "HA" at C3 — the do of do-re-mi — until it was muscle memory in my ear and my throat both.
“Without a ‘home note’, it's like navigating google maps without knowing your current location. You need one pitch you can reliably hit, even in the middle of the night. Once you do, you can navigate the rest of relative to it.”— Day 5 Video, Learning to Sing
From there you build outward: lock C3 with a pitch-meter app, fix wandering pitch with sirens, then walk intervals — DO, RE, DO — until the map fills in. It's the same instinct as everything else I make: find the one fixed point, then iterate off it. I feel strongly enough about this to have sketched a whole video around it — YOU ARE NOT TONE DEAF — aimed at people who think they're tone-deaf but actually just never listened to enough music growing up to build a sense of pitch. Most of them aren't broken. They just never found a home note.
The drill stack underneath it is real and a little obsessive — a twenty-minute warm-up plus an hour block, run with a tanpura droning the whole time. I know the jargon now, which still surprises me: lip trills are a semi-occlusive vocal-tract exercise, there's a straw-in-water drill, ZZZ and WWW in the lower range, vocal fry, soft-palate yawns, ear training on five-tone solfège. I even have a note on placement — the voice should sit right at or in front of the lips for a natural deep sound, not back in the throat.
- Warm-up: hot salt-water gargle, honey water, lip-roll sirens, "Tss" for breath stamina, yawn sirens, vocal fry, humming every note G2 to D5
- Solfège ladders — normal ladder, ladder jumps, then "minimal assist" where I shave the piano away while keeping accuracy
- Practice mapped line-by-line on "Can't Help Falling in Love": DO SO DO, ha, wise men say — with a margin note that the original is two keys up
- A standing reminder to myself, repeated like a mantra: KEEP TANPURA ON
It's a crooner's taste driving the song choices, honestly — Sinatra, Elvis, Billy Joel, Bublé, the Beatles. The goal was never range. It was one American song sung really well, and one Japanese song sung kinda okay.
Why I made singing worse on purpose
Here's the contradiction I'm proudest of, because it cost me something to choose. I optimize basically everything — I'll happily turn a hobby into a tier-list and a spreadsheet. With singing I did the opposite, deliberately, on the grounds that optimizing it would quietly kill it.
“It should not be geared towards actually learning to sing, even temporarily — that is just a blocker from actually accidentally learning to sing and having fun doing it long term.”— Monthly Goals Calendar
The rule it lives under is paint like no one will ever see the painting. So singing got reclassified from a learning session into break time — no pressure to improve, let alone to maximize improvement per hour. Just fun songs, fun note-phrases, rhythm games for the joy of them. It's near-daily now, about ninety minutes, a fixed checkbox in my journal that sits in my idea of a good day right next to my two hours of reading. Music does double duty as mood infrastructure, too — a "actually relaxing songs" set for the night routine, an energizing sing-along playlist I'm building for mornings — so the practice is wired into how the day feels, not bolted on as a task.
The whole trick was making it the lowest-stakes thing I do, so I'd actually keep doing it. That matters more for me than for most people, because my honest bottleneck has never been stopping — it's starting. I reject grind culture pretty firmly; the failure mode I actually fight is the blank first minute, not the discipline to continue. De-optimizing singing is me removing the one thing that makes starting hard: the pressure to be good at it. If you want the highly-optimized version of me, that's the Maker. This is the one room where I let myself be slow on purpose.
Standup: it's very important to be funny
I think being funny is a priority, and I'll say that with a straight face. My comedic voice is dry, structural, a little anti-authority, and built on a long-running bit about confusing people by quietly changing my own name mid-sentence and never acknowledging it. A sample of the actual material, which tells you the shape of it better than I can describe it:
“Okay first, name isn't Alex, don't call me Alex. Second, yes they will. … At first.”— Funny and Jokes for Videos
The setup before that punchline is a deadpan, escalating rant about burning every report card and never grading anything ever again — the joke is the bit where the imagined heckler calls me Alex and I correct them mid-argument without losing the thread. It's meta and self-referential, which is the only register I can do honestly. And I study comedy the way I study film and story: by hand, on purpose, refusing to outsource the part that's supposed to be mine.
“TODO: take one Drew Gooden video script and collect every single joke in it by hand (DO NOT USE FUCKING LLMS).”— Funny and Jokes for Videos
That last line is the AI rule I keep across everything I make: machines are welcome to do the labor, but the voice and the jokes are off-limits. A joke I didn't write isn't mine, and the whole point of being funny is that it's mine. So I do the unglamorous version — take one comedian's script and collect every joke in it by hand, the same critical-deconstruction habit I use on film and story, so I'm learning the machinery of a laugh rather than just stockpiling lines. The dedicated standup note in my vault is almost empty, which is honest: the ambition is real, but the actual work lives in the joke pile and the by-hand study, not in a tidy plan I haven't done yet.
Improv, voice, speaking — the rest of the kit
These are the threads that turn "funny once" into "reliable in a room." Some are done, some are prepped and waiting, and I'd rather be exact about which is which.
- Improv — past tense, and I want more. I've actually taken classes. I went for the unglamorous reasons: thinking, acting, speaking, and pure social-awkwardness reps. It's an underrated multi-tool, not a lark, and I'm not done with it.
- Voice acting — prep done, deliberately chosen over screen acting. The reasoning was that it's more specific and more useful — voiceovers, videos, narration — so I built the prep: daily cold reading recorded in one take, scripts marked like an actor with slashes for pauses and emphasis underlined, "one breath" takes, and a habit of watching and imitating people who talk for a living. I haven't performed it for real yet. That's honest, not modest.
- Speaking, as a curriculum. I broke communication into two halves I drill separately — Content (picking a theme in the first pause, organizing salient points, compelling an argument, generating examples on demand) and Delivery (rate of speech, melody, pauses, transitions, killing filler words, taking live interruptions). I even have a name for the kit — "core vocal charisma": rate of speech, volume, melody, facial expressions, pausing. The targets are specific to the point of being ridiculous: think at 300 words per second, build a real argument in three minutes, aim for zero "um." The single most useful delivery note I've ever written for myself, which I stole from how I run a D&D table, is to speak like I'm actually interested in everything I'm saying, like there's a very interested listener who doesn't know any of it — slow down, emphasize, care.
- Dance — light, and named honestly. Basics: a bit of social dancing, nightclub essentials, the bucket-list line of doing one TikTok dance perfectly. I'll also be upfront that part of the appeal is that it's a connection vehicle, which it is for most people who learn it.
The TEDx talk that's written and waiting
The artifact I'd point a skeptic to is a full TEDx draft — not an outline, a complete roughly-hour-long script, built around three pillars of a happy and successful life: sleep, health, and knowledge. It's built to work an audience, with engagement beats planned in the margins. It opens cold — Hi, I'm Swarnim, and what I'm going to talk about in this next hour will change your life. Forever — and runs through my own rock-bottom-to-recovery story before landing the closer I typed out months before there was ever a stage:
“When I said that the information in this talk was going to be life-changing, I meant it. Mike drop.”— Full Draft of TEDx Talk
The spine of it is my own growth-mindset story — the bit where I bottomed out early in college on no sleep and worse grades, decided to change exactly that, and watched it snowball into the opposite. It's the anecdote I reuse because it's true and because it earns the right to tell anyone else to fix their sleep. I have a stack of debater drills compiled to back the delivery — rapid-fire content generation, pencil-in-mouth articulation, no-filler runs where one slip means start over, speed "spreading." The usage rule is modest on purpose: pick two or three drills a day and rotate, rather than trying to do all of them and doing none. The thing I actually want is a month embedded on a competitive debate team, getting the reps I can't manufacture alone. (The talk is written; whether I've ever stood on a TEDx stage is a separate question, and I'm flagging it rather than implying it.)
Charisma as a built skill, and the part underneath it
The most validating data point I have isn't a stage. It's that in Japan, strangers kept assuming I was an actor or a model — and I made something like a hundred friends, about twenty close, in three months, including a few first conversations that ran five hours. I didn't get there on natural charm. I got there by treating presence as a design problem: study what makes someone feel trustworthy and put-together, then apply it systematically. Charisma, for me, is a built skill with a curriculum, same as singing — which is also why the place it gets the most reps isn't a stage at all but a D&D table, where for years I've run real-time collaborative storytelling, world-building, and crowd management for four-to-six people at once. The DM seat is where the improv, the voice work, and the structural comedy quietly compound; more on that side of it lives over in the worlds I run.
But the engine under all of it isn't a technique, and I'd be lying by omission if I left it out. The performance project grew out of a genuinely low stretch — a stretch where, at the height of money and luxury in New York, I mostly felt unlikable, like there was something I couldn't fix. The thing that actually worked wasn't a charisma drill. It was learning to enjoy my own company first.
“Liking myself made me that much more likeable.”— the load-bearing finding
That's the real order of operations, and it's why this page sits under "reinvention" and not "talents." All the drills are downstream of that one finding. The home note, the deadpan bits, the cold reads, the curriculum — they're the operational layer on top of a quieter discovery that the most attractive thing I could build was a self I didn't need an audience to validate. Which is, I'll admit, a funny destination for someone who writes hour-long talks.
What I've done, and what I want next
What I've done
Sang from genuine tone-deafness with real lessons and a public video; built the "home note" method and a daily, deliberately de-optimized practice. Took improv classes. Wrote a full TEDx draft and a body of dry, structural standup material studied by hand. Did the voice-acting prep, and built the kind of presence that had strangers in Japan assuming I performed for a living.
What I want to do next
A month embedded on a competitive debate team for reps I can't fake alone. Voice acting for real — prep into performance. A stage for the comedy, once the writing's earned it. And to keep singing badly, on purpose, forever — because that's the one I'd never want to optimize.
If any of this is your kind of thing — you run a stage, you do improv, you want a singing buddy who's deliberately not very good, or you just want to argue something out for the joy of arguing — say hi. The Bard exists to find people, so this part isn't optional. It's the whole point.