A wearable sensor glove instrument I've been building since 2014, from COSMOS at UC San Diego to Hélène Jabbour's ongoing thesis at AUB. Demo reel + arXiv reference.

A two-year Creative Labs project on unifying music and dance under movement, co-funded by the European Union and led by the Arab Fund for Arts and Culture (AFAC). The vision, the team, and what the two years look like.

The question is one of definition; and I love those. This one is loaded, especially now that there are so many of kinds of them. I’m not going to talk only about the artificial ones, and...

If I’m using an LLM to write, I first brainstorm some ideas, edit my essay, and if I’m too lazy to keep editing, I drop it into Claude, ask it to edit it, hoping it doesn’t...

My morning routine involves brainstorming out loud until I find an idea I haven’t thought about before. Then I’ll write it down, transcribe my voice memo if I recorded it; then, if I...

A short experiment where a piano video seeds a Game of Life whose column density becomes a live spectral filter over the piano's own audio. Framework payload: intelligence as a closed action–perception loop, made audible.

Classical information theory evaluates compression by symbol reconstruction fidelity. I argue this is the wrong metric. The purpose of information is to produce behavior: a compressed representation is good if the actions it induces are equivalent to those induced by the original.

There is only one consciousness in the world, and it is mine.This sounds like solipsism, a position that all respectable philosophers avoid like a plague. I’ll try to partially embrace it in...

We are really bad at predicting how new technologies will shape us. When cars arrived, people called them “horseless carriages”, only capturing what was absent rather than what was...

For centuries, we’ve built technology the same way: we understand the parts, predict how they combine in pairs, then engineer the whole. The steam engine, the bridge, and the transistor are all...

In what seems to be her first Substack post, Fei-Fei Li writes about spatial intelligence, arguing that the next big thing(s) in AI are world models. Yann LeCun who recently left Meta to focus on...

Science has always been in the business of compression. You observe the world, you find patterns, and you encode those patterns into representations compact enough to fit in a human mind, or on a...

The Vertically Integrated Project Program (VIPP) functions as applied research rather than traditional coursework. Students should embrace three core principles: learn by building, share progress within the lab community, and prioritize comprehensive documentation.

We live in an age where we need to learn how to be creative with machines. And it’s an age that hasn’t started recently. Creativity has always happened through various tools we’ve...

What is data science to me?It’s an interesting question, especially coming from someone who once cared so little about data. My undergraduate years in mechanical engineering were a combination...

The Puzzle of BoundariesWhen I first started thinking about intelligent machines, before I knew anything about machine learning, I stumbled upon what seemed like simple questions: What are objects?...

We live in what we call space and time. Or do we? Let me start by questioning this for a moment.We think of space and time as separate entities, but their orthogonality is far from certain. They may...

In Is Language Becoming a Science, I argue that we’re moving from writing to designing writers. More recently, I suggest the possibility of giving machines “selves”. While society...

So how do I use AI assistants? This seems to be a common question these days (especially spamming me on YouTube ads).While seemingly banal, the question surprised me when I started to seriously think...

The fundamental question “why do neural networks work?” has a deceptively simple answer. It has to do with emergence; a ubiquitous, yet poorly understood—and hard to...

We use the word understanding as if it’s a substance that minds possess. A more useful view is much simpler: understanding is a mapping between the external world and an internal world,...

After an extensive conversation and reflection with ChatGPT, I’ve decided to make it write the essay from its own “perspective”. I tried to provide it with as much...

Something is radically changing in how we work, how we communicate, write, and read. And I’m not sure how to feel about it. Some of it is exciting and some almost disgusting. The deeper...

Do I need to code this algorithm? Do I need to learn the theory or practice the application? Do I need to remember everything in that chapter? Do I need to understand what happens under the hood; or...

Boundaries have followed me for years. I remember the first time they stopped being invisible in 2013, when my master’s work forced me to stitch together two ways of seeing the phenomenon of...

The Music Intelligence Lab now offers a Vertically Integrated Project Program (VIPP). Through this initiative, students can earn course credit, participate in long-term research, and explore music and technology intersections.

What's music intelligence? Is it just AI for music? The term Music Intelligence Lab might imply artificial intelligence generating music, but intelligence here means the ability to adapt to one's surrounding through abstraction and problem solving.

How does a collection of living, breathing cells self-organize into much bigger intelligent beings? In this article, I argue that intelligence comes in many forms, characteristic of the diversity found in living and non-living organisms alike.

Our grasp of the world has evolved alongside our capacity to shape it. Language provided initial frameworks for abstract representation, but the Renaissance introduced a revolutionary concept: space and time could be quantified through mathematics.

How do we define a point? Because I've never seen one, have you? A point is an abstraction. It doesn't exist in reality; and yet, it's a surprisingly simple concept to imagine.

Modern science recognizes that purely deterministic equations cannot fully explain physical phenomena. Randomness pervades most measurable systems, requiring probabilistic approaches for accurate modeling.

Ideas emerge from the ceaseless storm of bursts in a complex network of cells. At least that's what many scientists believe. But the relationship between neurons and the abstract things they generate is wholly unknown.

Machine learning is everywhere. The basic principles of machine learning are easy to understand. In this tutorial, I'll cover linear regression to introduce general concepts.