Demonstrations¶
The demonstrations are the factory's proving ground and coverage contract: nothing ships that can't
be shown, honestly, inside the portfolio. Coverage is achieved by the portfolio, never by one demo
doing everything. Each demo composes capabilities at varying readiness — see the
Capability Maturity Matrix for what is live vs planned.
Illustrative — the D1 arc, "raw reads to a ranked candidate lead molecule."
| # | Demonstration | Exercises | Honesty flags |
|---|---|---|---|
| D3 | Tuberous Sclerosis Complex (flagship) | The whole factory on one child — most engines + agents + the platform layer | Gene therapy preclinical; everolimus real/approved; decision support; pediatric caution |
| D1 | Secondary genomics → novel drug molecule | Engines 1→2→3, MLOps, workflow composer, Chai-1 | Chai-1 planned; generated molecules are research candidates |
| D2 | Imaging → cardiac CAC → statin PGx → longevity | Engines 4 + 6, Pharmacogenomics + Biomarker agents | MAISI synthetic imaging research/QA only; decision support |
| D4 | Pediatric molecular tumor board | Engine 5, E1 germline, single-cell, fusion-first | Decision support; trial matching is informational |
| D5 | Oncology / CAR-T & biologic design | CAR-T agent, Engine 7 design suite, Chai-2 | Chai-2 gated (partnership access); designs are preclinical |
| D6 | Autoimmune program | Autoimmune agent, HLA, Pharmacogenomics | Research-/trial-use biomarker inputs; decision support |
| D7 | Neurology suite (Parkinson's featured) | Neurology agent: Parkinson's, Alzheimer's, ALS, MS, stroke | αSyn-SAA / NSD-ISS staging are research-/trial-use, not routine diagnostics |
The unit of impact¶
The flagship TSC program follows the arc the whole portfolio is built around — weight → compression → hope: the weight of a multi-system diagnosis, the compression of months of cross-specialty work into one governed afternoon, and honest hope (a real approved therapy today, an open bench for what comes next).
How to read the demos
Completion demos exist to guarantee coverage and serve specialist audiences; they need not all run live. Every demo keeps its honesty flags on screen, and all clinical output is decision support for a qualified clinician — never autonomous diagnosis or prescribing.