In recent years, a linguistic shift has swept through North African studies: Berber must be replaced with Amazigh. International organizations update their terminology. Academic departments follow suit. Diaspora activists frame it as decolonization, reclaiming indigenous names from colonial imposition.

In London, around 2017, I remember a headline presented as a victory: mixed-race births had overtaken white births. The tone was congratulatory, as if the old racial order were dissolving into something cleaner and calmer. But the headline was never the story. “Mixed” rarely hardens into a stable third identity; it works more like a hallway, and hallways lead somewhere. Colorism tilts the floor. Lighter skin and more European features move with less resistance through the places that decide how comfortable a life feels: hiring, housing, dating, the small relief of being assumed competent. Nobody has to announce it. A reward system teaches without speaking, and gravity does the rest. When ease keeps collecting at one end of the spectrum, people adjust. What begins as strategy becomes habit, habit turns into taste, and taste eventually pretends to be nature. The lineage does not vanish; it thins socially, until it reads as a charming family detail instead of a lived constraint.

What happens when pain stops being a signal of distress and becomes a form of currency?

To understand why former colonial powers still refuse to apologize, we should start with a simpler question: why did Holocaust accountability become unavoidable?

An official pushes a file to the top. Another buries a file at the bottom.

You have seen this exercise, or at least you have heard of it.

Tunisia wants to be developed, but it cannot answer one basic question: who lives where?

Can we extrapolate the Prisoner’s Dilemma to an entire society? Or, in places where “me first” is the default and kindness needs an explanation, have we already turned life into one?

Notes and reflections on “Thinking Strategically” by Avinash K. Dixit and Barry J. Nalebuff.

In Tunisia, there is a story we tell without even remembering when we first learned it. It drifts through childhood like a familiar breeze, part legend, part memory, part truth.

Mass deportations are coming to Europe. This is not a hypothesis; it is a trajectory.

We educated Tunisians speak three languages. We think this makes us rich. It makes us homeless.

“Semite” was never a race. It was simply a family of languages that included Hebrew, Arabic, and Aramaic. It was nineteenth-century European scholars who twisted that linguistic grouping into a racial fiction. Ernest Renan even placed Yiddish-speaking Jews closer to Aryans. The category was never coherent, only useful. Initially, this pseudoscience lumped Hebrew-speaking Jews and Arabs together, contrasting them against the “Aryan” ideal.

Across the continent, students still learn philosophy from a curriculum built elsewhere, for a world that did not imagine them as equals. At its center stands a European canon whose stars often mix real brilliance with open racism.

From the 12th century onward, knowledge flowed into Latin Christendom from Arabic‑speaking cities and beyond. To be admitted into European libraries and universities, it had to pass through a filter: Latinization.

The colony clarified everything. No pretense. The superior races have all rights over the inferior races. Just appetite with a flag. The Native Code (French: Code de l’indigénat) listed the methods: collective punishment, arbitrary seizure, forced labor. A manual for extraction. It moved through paper, officers, courts. Slow machinery.

It is a bitter irony that Germany, the nation that gave the world the DIN-Norm and standardized everything from paper sizes to industrial screws, cannot standardize the definition of its own people.

There is a grim irony in the Western anticipation of the “Second Coming.” Societies that sit at the apex of global power, consuming the majority of the world’s resources, are waiting for a Savior to return and rescue them.

Silence in the face of external violence is rarely just blindness; it is preparation.

The demand for “compatibility with Western civilization” relies on a linguistic sleight of hand. “Western” is not a geographic location. If it were, nations in the far East like Australia or New Zealand would not be included in the definition.

The adage “History tends to repeat itself” is often presented as a form of wisdom, a sigh of resignation to the laws of human nature. But when examined through the lens of power, the phrase loses its innocence. It begins to sound less like a warning and more like sarcasm.

Notes and reflections on “Weapons of Math Destruction” by Cathy O’Neil.

“Made in Germany” was once a vow. Today it is a costume.

The language of the Tunisian people is denied recognition, not only as an official language, but as a language at all. It is dismissed as “darija,” a mere slang. The central thesis of this analysis is that Tunisians should advocate for the official status of their own language. The common misconception that Tunisian is a dialect of Arabic is as historically and linguistically flawed as claiming that French or Spanish are simply dialects of Latin. Indeed, the linguistic distance between Tunisian and Peninsular Arabic exceeds that of recognized separate languages like German and Danish. This essay will deconstruct the baseless justifications for maintaining Arabic’s official status.

Introduction

Introduction

Introduction

Introduction

Introduction

The Small Multiples Revolution

Beyond the Basics

The Matplotlib Problem

In the previous post, we learned the fundamentals of web scraping and built our first crawler for movie data. Today, we’re taking things further by building a GitHub users crawler. But more importantly, we’ll explore when scraping makes sense, when it doesn’t, and how to do it responsibly.

Overview

Introduction

Consider a fraud detection system with 99.9% accuracy. By most standards, this would be considered exceptional performance. Yet if fraudulent transactions constitute only 0.1% of the total, a trivial classifier that labels everything as legitimate achieves identical accuracy while providing zero value. Worse, it creates a false sense of security that could cost millions.

The least squares method finds the best-fitting line through data by minimizing squared errors - the foundation of linear regression and machine learning. Instead of memorizing formulas, we’ll discover the solution by working backwards from what we want to achieve.

While exploring the mathematical foundations of machine learning, I discovered Professor Gilbert Strang’s MIT lecture series: Matrix Methods in Data Analysis, Signal Processing, and Machine Learning (18.065).