About

This is the long version. If you want the short one, the homepage is shorter.

Kabul, before

I grew up in Kabul. The Kabul I remember had Friday lunches that took four hours and a cousin who’d corner you and explain his theory about Iranian movies until you actually started caring. School was a mix of memorization, religion, and brittle English textbooks that taught you “I am a boy. He is a boy. We are boys” as if those were the building blocks of a language. My mother taught me to read Pashto out of newspapers because she didn’t trust the school to do it. My father had a small business and a dignified, exhausted patience that I’m only starting to understand now.

I want to be careful here, because Kabul is the kind of place that gets reduced into either a war story or a noble-suffering tableau and neither of those is the city I lived in. The city I lived in had traffic, weddings, a kid down the street who could rewire a TV with a kitchen knife, books, jokes, gossip, the usual. War was the weather, not the subject. Most of my early life was, in the texture of it, just life.

That changed in August 2021.

The evacuation

The fall of Kabul was on August 15th, 2021. I left on the 23rd. The week between those dates compressed about a decade of decision-making into a hundred hours.

I’m not going to give you a scene-by-scene of HKIA — there’s enough of that on the internet, much of it written by people who saw less than they claim — but here’s the part that mattered. We got to the airport perimeter. There were three Taliban checkpoints between us and the Marines. At the first one I had my phone in my hand with the wrong picture on the lock screen and I changed it without looking down. At the second one, an older Talib looked at my father for a long time and then looked away. At the third one, a Marine yelled, somebody pulled me over the wire, and I was on the wrong side of my own city.

The C-17 took off heavy. I had my backpack between my knees. The guy next to me was crying without sound, which is a particular kind of crying I’d never seen before. We landed in Qatar. They processed us at a base I won’t name. Then Ramstein. Then Fort Dix, New Jersey, where I saw snow for the first time in late November.

I was nineteen.

The first American year

The first year is mostly a blur of administrative survival. ESL classes, paperwork, food court at the base, the slow work of figuring out which English idioms were sincere and which were sarcastic. Americans say “How are you?” and the only correct answer is “Good, you?” which is a small linguistic trauma when you’ve been raised on Pashto, where “How are you?” is the start of an actual fifteen-minute conversation about your family, your knees, the weather, and whether you’ve eaten.

I started reading Hannah Arendt that winter. I had a copy of The Origins of Totalitarianism that someone gave me and I worked through it sentence by sentence with a dictionary. I didn’t get most of it. The parts I got were the parts that explained my last week in Kabul better than anything I’d been told by an adult in Afghanistan. That book is the reason I applied to college.

Bard

I got into Bard on a full scholarship. I picked Philosophy and Politics because Arendt was there, because Plato had been the first Western thinker I’d read in a language I half-understood, and because if I was going to do this American thing I wanted to do it with the people who’d thought hardest about what humans owe each other under bad governments.

Philosophy and politics, as a combined degree, gives you two skills. The philosophy half teaches you to read slowly and notice when an argument is doing rhetorical work — when somebody’s smuggling a premise past you in the form of a vibe. The politics half teaches you to look at institutions and ask who set them up, who benefits, who quietly bears the cost. Both of these are useful. They are also useful together in a way that most undergraduate education doesn’t quite achieve, because the philosophy keeps you from falling for the latest political ideology and the politics keeps you from disappearing into pure abstraction.

The course corpus that ended up mattering most:

I’m not going to pretend Bard was a steady on-ramp. I had a winter where I almost dropped out. I had a semester where my English failed me in a presentation and the room got quiet in a way I won’t forget. I had professors who took me seriously and a couple who didn’t, and the second group taught me as much as the first. I worked the dining hall. I tutored. I learned how American academia actually functions, which is to say I learned how to read a syllabus, which is more of a skill than people admit.

The code

I started teaching myself to ship software in my junior year. Not because I had a plan, but because I had a problem — a personal-organization system I needed to build for myself, since the off-the-shelf tools didn’t fit the way I lived between four languages and three countries.

I started in Python. FastAPI, then Pydantic, then async, then SQLAlchemy and Postgres and Docker. I went deep on backend before I touched frontend, which was the right order for me but probably not the right order for everyone. Eventually I added Next.js, then Tailwind, then Cloudflare Workers, then LLM orchestration, then multi-agent infrastructure, then a Tailscale-based fleet across two machines so I could offload heavy compute to a desktop while I worked on a laptop in a café.

That stack — backend + frontend + edge + LLM orchestration + multi-machine fleet — is what I now call “Sulimania.” It’s not a product. It’s a chassis. I run my own life on it: my school work, my translation pipeline, my newsletter, my customer discovery, my code research. It’s also the proof-of-stack for the AI consulting work. If a single operator with a refugee budget can keep an empire stack running with AI leverage, that’s a useful template for a lot of people who think they need a fifteen-person engineering team to start.

The thesis

My senior thesis is called A Compass to an Absurd Existence. It’s a synthesis of three thinkers — Nietzsche, Camus, Arendt — who otherwise don’t sit comfortably at the same table.

For a non-philosopher: Nietzsche says God is dead and we are now responsible for the values we used to outsource. Camus says yes, but the absurdity of life is not a problem to solve, it’s a condition to live well inside of — like a marriage, basically. Arendt says fine, but if we don’t pay attention to what happens between people in public, the void where God used to be gets filled by ideologies that will absolutely come for you. My thesis braids those three positions into a usable compass — not for “what’s the meaning of life,” because that’s the wrong question, but for “how do you act with dignity when the meaning isn’t given to you.”

That’s a refugee question. It’s also an immigrant-in-a-strange-country question, an exile question, a 21st-century question. I submitted the first version on April 29, 2026. Two more chapters are going out in late July. Excerpts will be on this site as they come.

The thesis sits on top of a stack of shorter papers that work the same problem from different angles. The Right to Die! (Berkowitz, 2025) is the Arendt-on-rightlessness paper — what does it mean to lose the political community that makes rights enforceable, and is the right to have rights an actual answer or just a beautifully-worded admission of defeat. The Right to Have Rights essay (same course, earlier) traces dignity from Roman dignitas through Christian universalization through the post-WWII human rights regime — and asks whether dignity is durable or whether it always collapses back into the cohesion of whatever political community is currently doing the protecting. The Philosophy of Language paper (Hagberg, 2025) is on truth in Rorty and Hacking and Wittgenstein — truth as something made in sentences rather than found in nature. Those three essays are the bedrock the thesis stands on.

Mosaic

I also cut tile. I’ve been making mosaic murals since 2023 — mostly geometric pattern work, occasional figurative pieces, the lineage somewhere between Herati kashi and the older folk-mosaic tradition that doesn’t get written about much. It’s the part of my practice that doesn’t render to a screen, which is exactly why I keep doing it. The full version of that work lives at /murals.

Family land

I own inherited family land in four Afghan provinces — Ghazni, Paktia, Paktika, Nangarhar. The deeds are qabaala-style (قباله), which means they describe the parcels by markers — a neighbor’s mulberry tree, a streambed that runs in the spring, a stone with a particular crack. There are no GPS coordinates. There is no satellite layer. The land exists in the memory of older men, most of whom are no longer in Afghanistan.

I’m slowly building a digital atlas of it using QGIS, satellite imagery, and the deeds I have. It’s a hobby project that will become something larger eventually. It’s also a small private war against forgetting — and I am aware, every time I open a layer in QGIS, that this is the same problem Arendt was naming. The land does not exist to a state that does not see it. Markers are not coordinates. Memory is not a record. Refugees know this in their hands.

What “Sulimania” means

People ask. The name is a play on my own — Suliman → Sulimania, like a country I’m building inside my own work. Half-serious. The deeper version is that I’m not interested in shipping one product, getting an exit, and going home. I’m interested in building a multi-division thing — translation, AI engineering, writing, advisory, eventually a small media operation, eventually maybe a foundation that does something useful for refugee education. The empire frame is unfashionable in startup culture, which is fine, because I’m not from startup culture. I’m from a culture that builds slowly across generations and doesn’t apologize for ambition.

The current operating model

I run on AI leverage. I am one operator. I have a paid model lane, a free model lane, a self-hosted local model on a desktop in upstate New York, and a routing layer that picks the cheapest one that’s good enough for the task. I have multi-agent infrastructure for delegation. I have Cloudflare Workers handling edge work. I have a Tailscale network connecting my machines. None of this is exotic. All of it is reliable. The point of it is that I can do the work of a small team without having a small team, which means I can take on translation gigs, AI consulting clients, school work, the newsletter, and customer discovery without dropping any of them.

It is, at the same time, exhausting. I want to be honest about that. The romantic version of “one operator with AI leverage” is a Twitter post. The real version is a calendar that reaches into Sunday and a body that needs more sleep than it gets. The leverage is real and so is the cost.

Refusing the inspiration-porn

I want to say this clearly so the rest of the site can be read in the right register.

I’m not here to make you feel good about America by being a grateful refugee. I’m not here to make you feel good about yourself by being an inspiring story. I’m not interested in being a mascot, a token, or a TED Talk. The refugee experience is real and the gratitude is real and the story is, by some measures, remarkable, but if any of that becomes the point of how I’m read, I become smaller than I am.

What I want is to be useful. Hire me to translate something. Subscribe to the newsletter. Hire me to build the thing your team can’t afford to hire someone full-time to build. Read the thesis when it’s out. Push back on something I wrote. That’s the relationship I’m offering.

The story is just where I’m coming from. It’s not the product.

— Sulim Annandale-on-Hudson, NY May 2026