My interest in this work begins long before I encountered it. Cycling has been a constant in my life... A way to stay grounded, present, connected to the physical world in a time when so much of modern life drifts into abstraction. I’ve built bikes, photographed them, collected images of them and raised my kids on two wheels. The bicycle has always been more than an object to me; it’s a form of thinking, a way of paying attention, a practice in meditation.

So when I first saw @mattdesl’ Latent Dispatch studies, the Bicycle drawing was an instant hit.

DesLauriers’ project interrogates how machines learn and generalise, how they smooth out the specificity of the world and converge toward archetype. The system he designed evolves single line drawings toward the statistically “pure” form of a thing. In the Bicycle study, that convergence is unmistakable. The machine finds the most recognisable outline: efficient, coherent, undeniably correct. Almost nothing is left extraneous.

And yet, that very correctness reveals a tension. Cycling has always been, for me, a conversation between precision and improvisation: the engineered frame versus the imperfect human body, the clean geometry of the machine versus the lived messiness of the ride. DesLauriers’ piece sits exactly in this same gap. Between the computationally distilled idea of “bicycle” and the lived experience of bicycles that resist averaging.

His recent Substack essay (link in 🧵) clarifies that Latent Dispatch is not a static project but the foundation of his PhD research into representation, entanglement and the politics of machine perception. These Studies capture the system before human interaction reshapes it. Before audience drawings feed back into the training data, before the machine begins to carry traces of other people’s gestures.

Owning the Bicycle study, then, is not just about acquiring a beautiful single line form. It is acquiring a pre entanglement artefact: a snapshot of the system before it evolves, before it absorbs the world around it, before the feedback loop alters what the machine understands a bicycle to be.

In a broader sense, this piece resonates with a cultural moment where AI increasingly dictates how things “should” look. The Bicycle study freezes that process at its infancy. The machine at the moment of certainty, before uncertainty returns through human intervention.

For me, this work sits at the intersection of personal history and contemporary inquiry. It reflects a lifetime spent around bicycles and a fascination with algorithmic drawing. But it also marks a pivotal moment in DesLauriers’ practice: the beginning of a research trajectory that will only become more complex as the system learns from us and we learn from it.

In that way, the Bicycle study is more than a picture. It’s a record of a system before it becomes entangled with the world... The machine draws the bicycle that belongs to no one. Your life is defined by bicycles that belong entirely to you.

My interest in this work begins long before I encountered it. Cycling has been a constant in my life... A way to stay grounded, present, connected to the physical world in a time when so much of modern life drifts into abstraction. I’ve built bikes, photographed them, collected images of them and raised my kids on two wheels. The bicycle has always been more than an object to me; it’s a form of thinking, a way of paying attention, a practice in meditation.

So when I first saw @mattdesl’ Latent Dispatch studies, the Bicycle drawing was an instant hit.

DesLauriers’ project interrogates how machines learn and generalise, how they smooth out the specificity of the world and converge toward archetype. The system he designed evolves single line drawings toward the statistically “pure” form of a thing. In the Bicycle study, that convergence is unmistakable. The machine finds the most recognisable outline: efficient, coherent, undeniably correct. Almost nothing is left extraneous.

And yet, that very correctness reveals a tension. Cycling has always been, for me, a conversation between precision and improvisation: the engineered frame versus the imperfect human body, the clean geometry of the machine versus the lived messiness of the ride. DesLauriers’ piece sits exactly in this same gap. Between the computationally distilled idea of “bicycle” and the lived experience of bicycles that resist averaging.

His recent Substack essay (link in 🧵) clarifies that Latent Dispatch is not a static project but the foundation of his PhD research into representation, entanglement and the politics of machine perception. These Studies capture the system before human interaction reshapes it. Before audience drawings feed back into the training data, before the machine begins to carry traces of other people’s gestures.

Owning the Bicycle study, then, is not just about acquiring a beautiful single line form. It is acquiring a pre entanglement artefact: a snapshot of the system before it evolves, before it absorbs the world around it, before the feedback loop alters what the machine understands a bicycle to be.

In a broader sense, this piece resonates with a cultural moment where AI increasingly dictates how things “should” look. The Bicycle study freezes that process at its infancy. The machine at the moment of certainty, before uncertainty returns through human intervention.

For me, this work sits at the intersection of personal history and contemporary inquiry. It reflects a lifetime spent around bicycles and a fascination with algorithmic drawing. But it also marks a pivotal moment in DesLauriers’ practice: the beginning of a research trajectory that will only become more complex as the system learns from us and we learn from it.

In that way, the Bicycle study is more than a picture. It’s a record of a system before it becomes entangled with the world... The machine draws the bicycle that belongs to no one. Your life is defined by bicycles that belong entirely to you.