Foundation AI for human motion

Large Movement Models

A foundation model that treats movement as a first-class sequence domain — learning the structure of how bodies move through the physical world.

Developed by LMM Technologies — an Aegis Station Infrastructure company
30%
lower motion-forecasting error vs. a standard transformer baseline
17k+
clips processed across five independent capture sources, zero pipeline failures
5
capture modalities converge to one normalized representation
<150MB
model + pipeline footprint — runs on a single GPU, fully air-gapped
The frontier is moving

Foundation models proved themselves on text. The race now is the physical world.

In late 2025, Jeff Bezos took his first operational role since Amazon to co-found Project Prometheus — an AI company, since valued near $41B, built to make models that understand physical reality rather than just language. It is the loudest signal yet of a broad turn: the next era of AI is embodied.

Language was the first domain to fall to the foundation-model recipe because it was the easiest to digitize. It will not be the last. Any domain that unfolds in time and carries hard structural constraints — markets, biological signals, industrial telemetry, and above all movement — is a candidate for the same treatment.

LMM Technologies builds the foundation model for one of the most fundamental physical signals there is: how a body moves through space. Not a CAD assistant, not a robot — the motor prior underneath all of them.

The thesis

Machine learning, or manufactured instinct?

We are not building artificial minds. We are manufacturing artificial instinct — compressing the structure of a domain into priors the way evolution did for living things, only on a vastly shorter timescale.

Priors, not reasoning

Instinct is compressed structure

A foal stands within hours; a trader reads a tape at a glance. Expertise lives in priors built from exposure, not deliberation. Large models manufacture exactly this kind of prior.

The recipe generalizes

Movement is a sequence domain

The same recipe — large-scale training, learned tokenization, attention over long sequences — applies wherever an expert sees what a novice cannot. Motion is a prime case, and a largely untouched one.

Instinct first

The first half of an architecture

Today's models are the bottom layer: manufactured instinct. The half still missing is learning — adapting from little because it stands on a great deal.

Read the essay →

How it works

From raw video to motion intelligence

A Large Movement Model treats motion itself as the primary data type — sequences of joint positions, trajectories, and timing — and learns how movement unfolds the way a language model learns how text unfolds.

01

Capture

Video or sensor streams from any camera, frame rate, or skeleton format.

02

Pose

OpenPose BODY_25 keypoints — 25 joints with per-joint confidence.

03

Normalize

Hip-centered, torso-scaled, resampled to 15 FPS with QC scoring.

04

Tokenize

Frame, window, and body-part views — motion as multi-resolution tokens.

05

Model

Hierarchical transformer or diffusion inference over the sequence.

Multi-resolution by design. Movement happens at many timescales at once — frame-level dynamics, half-second phrases, whole-body coordination. LMM attends to all three simultaneously, which is what lets it stay coherent seconds into the future where flat models drift.

See the full technical overview — architecture, results, and cross-domain transfer →

Where it applies

One motor prior, many domains

The pipeline makes no assumption about what kind of movement it sees. In Phase I, motion dynamics learned on dance video transferred — essentially intact — to clinical motion capture the model had never seen.

Rehabilitation

Movement-quality analytics, recovery tracking, and compensatory-pattern detection across gait and balance.

Sports & performance

Technique assessment, fatigue signals, injury-risk estimation, and biomechanical pattern analysis.

Robotics & embodiment

A pretrained human-motion prior for imitation learning, retargeting, and anticipatory human–robot coordination.

Ergonomics & safety

Workplace movement monitoring, physical-demand quantification, and real-time anomaly detection.

Behavioral analytics

Gait recognition, anomalous-movement detection, and intent inference from body language.

Human–computer interaction

Motion-aware interfaces that read intent from how a person moves, not just what they touch.

Embodiment

A motor-cortex prior for any body

LMM internalizes that shoulders have range-of-motion limits, that gait is periodic, that balance demands continuous postural adjustment. This knowledge exists independently of any particular body.

Most humanoid and assistive robots are trained from scratch in simulation — expensive, with brittle sim-to-real transfer — or rely on hand-coded primitives. A model that already understands joint coordination, plausible trajectories, and multi-timescale dynamics gives a robot a head start: fine-tune from a rich prior instead of learning from zero.

The architecture lines up with the problem in a way that's hard to ignore: the HTT's three temporal levels correspond to a robot's actuator control, trajectory planning, and task sequencing, and the body-part stream is the kind of coordination signal that bimanual manipulation and whole-body balance demand. In principle, a 1–2 second forecast horizon is the right window for a robot to anticipate a human collaborator rather than merely react. Whether the motion prior transfers this way is a direction we're pursuing — not a result we're claiming.

Go deeper

Read the work

Aside — off the movement beat

Preparing for the Data Apocalypse

A change of pace. When the grid goes down and "just Google it" stops working, the most valuable item in your kit might be a local AI model — on building an offline knowledge archive.

Read the dispatch →
Get in touch

Building the foundation model for movement

For research, partnership, or pilot inquiries across rehabilitation, sports, robotics, and embodied systems.

contact@lmmtech.ai