Arjun
Dhall

I am not going to write a cleaner origin story than the one I have. A specific detail I have not checked is worse than a blank, so the blanks stay until I fill them in my own words.

The through-line I am willing to claim is ordinary, and I think it is enough. I like taking something confusing — a system, a product, a business, a model — and staying with it until the mechanism is visible. Building is one way I do that. Reading how a company actually makes money is another. AI has become a third, mostly because it changes how fast I can test an idea, and how careful I have to be about what I trust.

I like building because a sentence can sound right and still fall apart once someone has to use the thing. I would rather find that out on a rough version than after a long plan.

[ADD PERSONAL DETAIL]Where you started. A place, a school, a household, a first job — only what is true.

What I cannot sketch from the outside is the path. The order matters, and it has to be mine: what pulled me toward technology, what work changed my standards, and what later made business, markets, or AI feel like part of the same question.

[ADD PERSONAL DETAIL]How your interest in technology developed, and a professional experience that changed how you work.

[ADD FINANCE/INVESTING INTEREST]Where your interest in business, finance, investing, markets, or AI actually comes from.

Next — What I build

A list of tools does not tell anyone how I think. I would rather take a few things I actually made and say why they existed: the problem, why I cared, what I built, what was hard, what I learned, and what I would change.

Three places are here. I would rather describe three properly than imply a portfolio I have not written up. The technology can show up inside the story. It should not be the story.

Next — How I use AI

I do not think the interesting question is whether someone uses AI. The tools are already on the desk. The interesting question is where I let them move fast, and where I slow down.

I use them to get to a rough version sooner. A first pass at an interface, a script, a way into a subject I do not know yet. The draft exists so I can react to it. It does not exist so I can skip the reaction.

I get more careful when the output is a fact, a number, or a claim about a real person or a real company. Those are easy to phrase and tedious to verify, which is why they get waved through. Fluency is not the same thing as evidence.

I am more interested in a product where a person can tell whether the model did the job. If the user has no way to know the answer is wrong, the demo is ahead of the product. There is a matching skill on the other side: sometimes a spreadsheet, a query, or a plain program is the better tool, and a model would only add a new way to be confidently wrong. I want to get faster at telling those cases apart.

That is also why AI changes how I learn, not only how I type. I can cover more ground in an evening. The risk is that coverage feels like understanding. It is not, until I can explain the mechanism without the model in the room, or catch the model when it smooths over a hole.

[ADD RECENT LEARNING]A specific time you used AI to build or learn something. Where it helped, and where it misled you.

Next — What I'm thinking about

These are not a syllabus, and they are not claims of expertise. They are the questions I want this site to be honest about. If one of them is not really mine, I would rather replace it than perform it.

[ADD RECENT LEARNING]Replace any of the questions above with what you are actually spending time on, in your words.

Next — Businesses and markets

I should be plain about the level I am at. I am learning. I do not have a record as an investor, and I am not going to borrow one. What I have is a set of questions I want to be able to answer without hiding behind vocabulary.

A business can look wonderful in a story and ordinary once you look at how cash actually moves. I want to be the kind of reader who notices which of those I am looking at. Durability is the question I keep circling. Something can be popular and still be easy to leave. Something can be quiet and still be painful to replace. I do not yet have a reliable feel for which advantages survive contact with a competitor, a new tool, or a customer who is simply busy.

The story

How cash moves

Technology makes that harder, not easier, because it tempts people to treat a product demo as the business. Cheaper software changes the cost of building. It does not automatically change distribution, trust, switching costs, or the reason a customer stays. I am trying to separate those.

AI companies feel like a particular trap for loose thinking. Usage can grow while the economics still depend on someone else's model, someone else's distribution, and a customer who can leave. I do not have a finished view of how to read that. I do have a suspicion that the usual software comforts — high gross margin, a sticky product, the rest will follow — are not enough on their own.

The other question is incentives. Inside a firm or in a market, people respond to what they are paid to notice. A decision can look sensible up close and expensive later because the person closest to it was never asked to carry the later cost. I want to get better at seeing that before I am impressed by the narrative.

[ADD FINANCE/INVESTING INTEREST]Anything you have actually done or studied: a class, a company you followed, a model you built, a book that changed how you see a business. If you have not done these yet, say what you want to learn first.

Next — How I work

These are working notes, not a creed. I break them. They are still the standard I want to be held to.

Next — Experiences

A list of titles would be the wrong shape. I want the moments that changed how I work: real users, a skill I had to teach myself, a room where I had to explain the work to someone who did not share my vocabulary. I have not written those yet. The prompts below are the ones worth answering.

Next — Why Recalc

Recalc is interesting to me because of the order it insists on. Learn the mechanics of a business: how to read it, how a model is put together, why a number means what it means. Use AI to move through that work faster, not to step over the understanding.

That is close to the way I want to build. A quicker first version is a good thing. An explanation I cannot defend is not. I want those two standards in the same place, aimed at companies and markets, not only at software.

I also want the room. People who will argue with a thin assumption, who are serious about getting better, and who are thinking about technology, businesses, and capital at the same time. I do not mean a network in the abstract. I mean a group I can learn from, and be useful to.

What I can offer is a working habit, not a finished point of view on finance. Give me a messy problem and I will try to make a first version, then test whether it holds. When AI is in the loop, I still want a person responsible for what gets kept. I am early in finance. I am not trying to hide that. I would rather learn it in a cohort that expects the work.

[ADD PERSONAL DETAIL]One specific reason Recalc fits your actual situation right now. Timing, a question you cannot answer alone, or the kind of person you want to learn from.

Next — The short version