About
Maximize humanity's long-term ability to understand, build, survive, and flourish.
Four verbs, four failures ruled out: ignorance, helplessness, extinction, mere persistence.
2026
Now
Software only. Zero physical hardware.
2026 to 2035
Prototypes and dead ends
Build some ideas, drop the rest.
~2035
Formal launch, target
A current working plan, not a promise.
- Target, not a promise
- Most of this map will not exist by then
- Not a funding round, not a launch date
How early this is
- Nine branches. Areas on a map, not staffed departments.
- Nobody is staffed to any of them.
- Four software projects have work happening on them.
- ME has not launched and is not trading.
What exists, and what does not
- BUILDING NOWCommits this month
- PLANNEDNext step written down, not started
- LONG-TERMDirection only, nobody building it
- VERY LONG-TERMAmbitious idea, nowhere near
Only BUILDING NOW claims work is happening. The other three do not exist.
- ME OS
- Boots in QEMU. Never on physical hardware.
- Holoprojector
- Drives a simulator. No projector exists.
- Carl
- One personal machine. Deployed nowhere.
- Physical hardware
- None yet. Not one prototype.
The roadmap says how to check each one.
Why the map is wide
Where ME could go, not what is being built. Early work makes technology later work reuses.
- Blockers are elsewhere: a display needs materials
- Results transfer: wearable sensing feeds prosthetics
- Understand, build, survive, flourish is not one discipline
- ProductsHoloprojector · Bracers · Lab Suit · Driver
- Agent layerCarl
- Shared toolsResearch tools · Manufacturing
- Computing baseME OS
Broad roadmap, narrow active scope. A project goes active when its next milestone is small enough to show.
Before any launch
A portfolio, not a product
Software prototypes, concepts, research. Hardware only once a question needs a physical answer.
Foundations first
An operating system, a command path, calibration, safety checks. Every later branch borrows them.
Capabilities that get reused
Routing for a projector is what a robot needs.
Honest status labels
Every product carries a tier. ME has no physical hardware prototype, and no page says otherwise.
Current hypothesis
Carl running on ME OS as an agent native computing base
What should ME be able to show in three to five years?
Not a device. The agent layer paired with the operating system built to host it, with a human override.
- Both projects are already active
- Every other branch wants the same base
- Software, so it needs no factory
A hypothesis, not a closed decision. Kept open so it can be argued with.
What this is not
- No investors, no funding, no valuation
- No customers, no orders, no revenue
- No partnerships, no suppliers, no contracts
- No product for sale, no ship date
- No physical hardware prototype
- No employees, no departments, no staffed branches
- Not launched. Around 2035 is a target, not an event
Interfaces should be multimodal
One channel suits one kind of body on one kind of day. None of these exists.
Wearable and muscle sensing
ResearchBracer tracking, muscle sensing as a question
Eye tracking
ConceptGaze as pointing, where hands are occupied
Voice
ConceptSpoken instruction on the same command path
Switches and adaptive controllers
ConceptExisting hardware, not ME specific hardware
Brain computer interfaces
ResearchA long horizon interest, outside anything ME builds
- One command path, so a new input rewrites no application
- Calibration at first use, then continuous adaptation
- Text, contrast and motion settings respected
- No claim that any ME system diagnoses, treats or assists with a medical condition
Public and internal
- This site
- Public project pages and an embedded authenticated chat portal.
- Internal systems
- AOS runtime controls remain separate from the public website.
Criticism is welcome. Tell ME what is wrong with this.