Machine Spirits — About

Machine Spirits — About

Who, what, and why. Started 2026-05-16.

The argument is in MANIFESTO.md (the Founding Sketch). The name, tagline, and voice are in IDENTITY.md. The full research program and its checkpoints are in STRATEGY.md. This page is the short orientation.

What this is

Machine Spirits is a research program: a thesis about AI and a growing body of work — software, teaching, and writing — advancing it. The platform does two things at once. It builds systems that perform an alternative account of machine and human learning, and it writes, in public, about what those systems do and fail to do. Critique and construction, each load-bearing for the other.

The tagline — Learning (to live) with Machines — is meant to be read both ways at once: learning with machines (a claim about how cognition forms in a shared field) and learning to live with machines (an ethical, everyday one). The program refuses to separate the two.

Why — the thesis

Contemporary AI is impoverished by a positivist, behaviorist, thin-ethics frame. Phenomenological, psychoanalytic, and social-theoretic resources — recognition (Hegel, Honneth, Benjamin), the unconscious and the nonconscious (Freud, Hayles), charismatic authority (Weber), temporal orientation (Heidegger via Pippin) — give a deeper and more workable account of human learning, machine learning, and the relation between them.

The program’s distinctiveness is precise and statable: recognition theory + working software + pedagogy, integrated. The philosophy is not decorative commentary on the engineering, and the engineering is not a mere illustration of the philosophy. Each is evidence for the other.

The name

“Machine Spirits” is doing several things at once:

• a Hegelian wink at Geist — spirit, both individual and objective;
• a Heideggerian wink at the uncanny — unheimlich spirits in machines;
• a Vygotskian wink at the cultural-historical formation of cognition;
• a colloquial wink at “ghost in the machine,” while explicitly refusing the Cartesian dualism that phrase encodes.

It is meant to be evocative to a non-philosopher and exact to a philosopher. That double legibility is the house style.

Who

Machine Spirits is built by Liam Magee at the University of Illinois Urbana-Champaign. The teaching limb — courses and course-notes — is the program’s institutional home; the software and writing limbs operate from it.

The platform is about the work, not a personality. Public artifacts are dated and revised in the open; the reasoning is meant to be followable, not announced.

How to cite

Cite the platform as:

Magee, Liam. Machine Spirits. University of Illinois Urbana-Champaign. https://machinespirits.org

Individual essays, probes, and reading-room dispatches carry their own dates and stable paths — cite the specific artifact and its date. Working documents are revised in git, and the date fixes the version you read. A CITATION.cff travels with the software repositories for tooling that consumes it.

Contact

Liam Magee — liam.magee@gmail.com. Social handles are listed in the site footer and in IDENTITY.md.