In 2017, I wrote my LSE thesis using computational text analysis. I ran parliamentary speeches through code to explore gender differences in the House of Commons. It was a niche method then. Most people I spoke with, almost everyone outside academia, had never heard of it. Prior to that, I’d been exploring the history of political thought in my undergrad. I didn’t expect the two to meet so drastically in my working life that followed. They did.
A decade in the room
I’ve spent ten years inside the technology. From policy to AI to blockchain. What’s stuck with me most is the meetings, from boardrooms to product mapping, systems design and schema reviews. Every technical trade-off carries someone’s philosophy. Speed vs accuracy. Cost vs privacy. What a system should store and ignore, and who gets to decide. It’s rarely (never) called philosophy, but philosophy and moral ethics power every requirement, they ship with every product.
Building and evaluating data and economic models early in my career taught me this best. Before any query is run, you’ve implicitly (or explicitly) decided what kinds of things exist and how they’re “allowed” to relate. Kant wrote about these structures that shape experience before it happens in the 18th Century. I kept seeing it in review meetings centuries later.
What changed
When I wrote my thesis ten years ago, text analysis essentially meant counting and classifying human words. Today we have models that write the words. This has raised a question for me I can’t put down. If what a language model produces isn’t human language in the usual sense, what is it?
Beatrice Fazi argues that these systems perform a kind of synthesis. Pulling scattered patterns into a unity, building a “representational world within”. Others push back. Pope Leo XIV’s encyclical Magnifica Humanitas argues that no machine can replace a person’s inner life. Either way, these AI outputs now shape what millions of people read, believe and decide.
Trust is shifting too. For most of history humans trusted in institutions to tell us what’s true. Not necessarily governments, but courts, banks, universities, newspapers etc. When we’re unsure whether the output we’re reading comes from a human voice or not, trust erodes, and people search for something else to trust.
Having researched and worked in web3, I’ve watched one answer to diminishing trust played out at scale. Blockchains try to move some of that trust into protocols, oracles and definitive “democratised” rules. Written in code and applied to everyone equally. Kant wanted moral rules that could hold for everyone. Smart contracts try something similar in code. I think he’d have recognised the ambition, but I’m less sure he’d like the result.
Why now
Philosophy has been my personal passion from a young age. A naturally inquisitive child, I constantly asked “Why”. I started reading early and naturally progressed to books on thought, power and how our world works. It’s why I studied politics and started my career in policy.
I’ve kept reading and writing the whole time. Notes on my phone. Books, papers and journals on the train. A folder of PDFs so big I probably do need AI to organise it.
This year, I’ve been slowing down and asking what I actually want. Time in between jobs allowed me to return to some of these questions at length. More time with these questions left me wanting more people to think them through with. As I did with coding in the open six years ago, I’m now giving these thoughts a public place to go.
What I’ll be writing about
Philosophy and computational systems. A broad academic phrase. For me it means:
- Synthesis and the world of machines. What LLM outputs really are, and how they mediate our reality.
- Structure before meaning. Ontologies, schemas and categories. What they decide before anyone asks a question.
- Trust as infrastructure. What happens when truth and trust move from institutions to networks and protocols.
- The philosophy in the room. The assumptions we build into systems and products without noticing.
- The long history of computation. My nerd space. From Leibniz to Turing to ELIZA. None of this is as new as it looks.
On X I’ll be posting short takes and reading notes. Here you’ll get longer versions when an idea needs more room or I’m inspired to write.
What this platform is
I’m writing as someone who has built (with) and evaluated this technology for a decade and still finds it strange. I may change my mind. Not every post will be fully formed. I’m committing, again, to learning in the open. Thinking out loud.
If you think about these questions too, I’d love to hear from you. Philosophers, data scientists, builders, people in academia and industry. Self-appointed professional nerds. Send me a paper you can’t stop thinking about, or a book I should read. Find me on X @rkulidzan or send me an email.
Let’s think it through together.