
When we think of "artificial intelligence" and "content generation," we often focus on text, images, or conversational responses. But the roots of this phenomenon are deeper and older than they might seem. And, interestingly, an important part of this story begins with music.

## 🎶 The origin: music written by machines

Algorithmic music is not a 21st-century invention. Back in the 20th century, composers like Lejaren Hiller and Iannis Xenakis explored musical creation through mathematical rules and algorithms. This approach wasn't just about generating random melodies, but about exploring new expressive paths where calculation replaced intuition.

Unlike improvised music, algorithmic music aimed to detach from the creative self and give way to pure abstraction: what would a symphony created by a rule sound like? The challenge was (and still is) to find meaning in something generated by a process that, in theory, has no emotions or intentions.

## 🐒 The infinite monkey metaphor

The [infinite monkey theorem](https://en.wikipedia.org/wiki/Infinite_monkey_theorem) states that if a monkey randomly presses the keys of a typewriter for an infinite amount of time, it will eventually produce any conceivable text—including the complete works of Shakespeare.

This idea, as absurd as it is unsettling, is key to understanding the heart of many current algorithmic generation technologies: it's not about understanding, it's about statistics. When a system generates a sentence, it doesn’t “know” what it's saying. It simply calculates which word is most likely to come next.

## 🤖 From randomness to language

Language models like ChatGPT don’t work exactly like a monkey hitting keys, but the underlying logic is similar: to produce coherent text based on statistical patterns extracted from a massive training corpus. The result may seem magical, but it’s pure computation.

Now, if systems generate more and more content online... what happens to quality? If the internet fills with AI-written texts trained on other AI-generated texts, could we [end up in a loop where originality vanishes and repetition reigns](https://aeon.co/essays/who-needs-ai-text-generation-when-theres-erasmus-of-rotterdam?utm_source=rss-feed)?

## 🧠 Between automatism and the unconscious

The idea of writing without consciousness is not new. In the 20th century, the Surrealists practiced **automatic writing** as a way to access the creative unconscious without rational filters. André Breton, Artaud, and Dalí wrote by letting thoughts flow, without censorship or predefined structure.

It’s not so different from how some generative systems work: they repeat patterns, produce unexpected outputs, and sometimes create works with a strange kind of charm.

## 📚 And what does Horace Walpole have to do with all this?

In the late 18th century, English writer Horace Walpole published [*Hieroglyphic Tales*](https://isbndb.com/search/books/Cuentos%2BJerogl%C3%ADficos), a collection of satirical stories full of absurdities, wordplay, and semi-narrative nonsense. Some critics have seen in this work a kind of "proto-text generator": a literary attempt to explore the chaos of language and meaning when rules aren’t clear.

This approach—playful yet critical—makes us think: to what extent can algorithms be creative? And we, as readers, how do we identify what has value amid so much statistical repetition?

## ❓What now?

We are facing a silent revolution. Millions of texts, songs, and images are being generated by systems that don’t think but mimic thought. This forces us to rethink creativity, ownership, and even the meaning of culture itself.

- Can we consider a melody generated by an algorithm as art?
- Does it make sense to write a poem if you can generate a thousand?
- How do we distinguish noise from the human voice?

And perhaps the most unsettling question: **if reading no longer means listening to a person, but to a system... who are we really listening to?**

---

> ✍️ This article may have been generated by an algorithm. It’s unclear whether what it says is true, but it definitely sounds coherent.


From algorithmic music to ChatGPT: how AI generates content, what the infinite monkey has to do with it, and what questions it raises about creativity.

From Algorithmic Music to ChatGPT


# The Project  
**Version 1.0**  
✍️ AI-assisted: This article was written with the help of ChatGPT (OpenAI), based on original sketches and ideas.

Over the past month, I’ve immersed myself in a personal project that was meant to be a game, an experiment, and maybe a bit of a protest: **WordGuardian**.

A site for playing with absurd, real, and invented definitions. But also a window into the semantic richness of Wikidata and a cold reality check about building an indie website in 2025.

## 🧩 Objective  
I wanted to create a game where users guessed plausible definitions, either automatically generated or extracted.  

**But the journey was anything but simple.**  
I used Astro and Preact to avoid FOUC and keep it responsive, and I worked with massive Wikidata dumps, decompressed with `pbzip2`, `zstd`, and parsed using parallel **Rust** scripts.

## ⚙️ Stack & Techniques  

- **Frontend**: Astro, Preact, Shoelace, Vanilla Extract  
  > *Astro turned out to be a hidden gem.* It lets you combine components from different frameworks and enables efficient **server-side rendering (SSR)**. This helps search engines index the content properly —a key difference if you care about SEO in 2025. Also, exposing a small server API (via `/api/xyz` endpoints) is simple and clean, no need for a full backend.  
- **Backend**: Node.js for scraping; parallel Rust scripts for parsing Wikidata  
- **Processing**: Wikidata dumps using p31, p279, p1629... a whole world of data  
- **NLP**: distractor generation with Groq, heuristics using embeddings.
- **Other**: Load without *FOUC* (*Flash of Unstyled Content*) —making sure the page doesn’t flash ugly before styles kick in. Minimal optimization, and as always, lots of console logs.

I also relied on the [**Wiktextract**](https://github.com/tatuylonen/wiktextract) project by **Tatu Ylönen**, and the dataset from [**Kaikki.org**](https://kaikki.org), which offers Wiktionary content in a structured, usable way. Without it, building a solid lexical base would have taken much longer.

## 💡 What I Learned  

- **Wikidata is a treasure**, but navigating it without a map is a nightmare: items, properties, lexemes, glosses, cross-references... and **SPARQL** queries that look like arcane rituals.  
- Tools I discovered:
  - [Wikidata Toolkit](https://github.com/Wikidata-Toolkit/Wikidata-Toolkit)  
  - [KGTK Search](https://github.com/usc-isi-i2/kgtk-search)  
  - [Wikibase CLI](https://github.com/maxlath/wikibase-cli/tree/main)  
  - [wikidata-filter](https://github.com/alexkreidler/wikidata-filter)  
  - [SPARQL](https://en.wikipedia.org/wiki/SPARQL)
- Wikidata dumps often come in **JSONL** (JSON Lines) format: a file where each line is a standalone JSON object. Great for streaming —you don’t need to load everything into memory. But with 90 GB... even the best strategy requires patience.
- **Compression formats matter**:  
  - `bz2` is common in official dumps, but *painfully* slow to decompress.  
  - `zstd` is **much faster**, especially with multiple threads. With `pzstd -d -p10` you can speed up decompression, although your own code becomes the bottleneck 😅
- Generating plausible distractors is not trivial. Doing it with Groq was fun but inconsistent.  
- Making fake but believable definitions is tricky. You need precise prompting with Groq or HuggingFace.  
- **Dumps are not for the faint of heart.** Especially when they’re 90 GB.
- I also learned to wrap small NLP services in Docker, exposing local APIs with ease. Using a lightweight image like `python:3.11-slim` and a `requirements.txt`, I quickly deployed a service for semantic comparisons and phrase similarity. Perfect for local tests without dependency chaos.
- I explored [**Common Crawl**](https://commoncrawl.org/) to URLs from the public web. Although I didn’t end up using it, I built a script to filter the links and create a custom corpus. The idea was to use it to generate or validate questions with real content. It’s on hold... for now.

## 📉 Monetization? Yikes...

I had hoped to at least cover the hosting with ads. But the numbers speak for themselves:

- 462 visits  
- 22 active users  
- €0.00 revenue  
- An average of 7 minutes per user... but no clicks

> In 2025, if you don’t have TikTok, a newsletter, or a million page views, ads won’t even buy you coffee.  
> Making a website isn’t enough. You have to be a showman, have money, or do daily marketing.

## 🎯 So, what now?

I’m not killing it.  
**WordGuardian will live on** as an experimental project and a small tribute to words and to Wikidata.  

Maybe it will evolve into an educational app. Maybe it’ll become the base for a new game.  
Or maybe it’ll stand as a monument to the time spent on an idea that still makes me smile.

I’m also toying with the idea of **exploring how to use Wikidata’s ontology to generate questions by topic** —science, history, pop culture, etc.  
Using subclass (`p279`) and instance (`p31`) hierarchies to build *thematic levels* or *topic-based quizzes*, making the game a bit less “psychedelic” and a bit more navigable.  

👉 Try it: [https://eurekatop.com/wordguardian](https://eurekatop.com/wordguardian/en/game)  
📖 Open source (coming soon on GitHub)  

A project by [**mutiitu.com**](https://www.mutiitu.com)

© Francesc López Marió, 2025


Over the past month, I’ve immersed myself in a personal project that was meant to be a game, an experiment, and maybe a bit of a protest: **WordGuardian**. A website for playing with absurd, real, and invented definitions. But also a window into Wikidata’s semantic richness and a cold shower on the reality of building an indie web in 2025.

Category: technology