Est. 2017

DMT SYSTEM

A framework for creating powerful decentralized realtime apps.

Local-first, entirely internet-enabled. A network of sovereign personal devices that makes everything around you just work better.

What it is

A local-first life assistant, running on machines you own.

DMT SYSTEM utilizes a network of sovereign personal devices to serve user-friendly distributed apps — so you can be faster and more organized, with less friction on everything that can be automated and monitored.

Control your surroundings

Home automation is the visible one — lights, audio, sensors, screens, doors. It is not the only one. Anything with an address on your network becomes something the system can reason about.

Get updates from the world

The outside comes to you, filtered and on your terms, instead of you going out to fetch it from a dozen tabs.

Automate the tedious

Periodic tasks get delegated to machines that do not get bored or forget. You get the attention back.

Monitor everything

Nodes report state continuously. When something stops, you know — and usually why.

Barrier to entry

Some hacker's knowledge. This is a tool for capable enthusiasts — developed and tested over ten years. It was almost impossible to develop from scratch; and the effort it took should not be underestimated.

The devil is in the details — and in the design choices.

Code with substance

Run DMT

We begin where computing began — in the terminal.

~/.dmt
  1. git clone https://github.com/uniqpath/dmt.git ~/.dmt
  2. cd ~/.dmt && ./install

    then reload the terminal

  3. dmt run

    dmt-proc liftoff

  4. open http://localhost:7777

    check that everything works so far

Where it runs

One engine, replicated to every machine you own — working for you 24/7.

Runs on

  • LinuxDebian, Ubuntu, Arch
  • macOSIntel and Apple silicon
  • Windowsvia WSL
  • Raspberry Pi OSPi 3 and newer

Deployed to

  • Personal computersyour daily driver
  • Serversalways-on nodes
  • Single board computersat the edge

Fundamentals

Four pieces

The system emerges out of multiple interacting DMT processes.
The more processes the better — but just one is enough to start.

  1. 01

    DMT Engine

    the process

    An evented distributed computing system which seamlessly performs useful tasks for you. Start one DMT process on your personal computer, then replicate the engine to every other device as you need it.

  2. 02

    Connectome

    the edges

    Connectome forms synapses and passes signals between DMT nodes. Built exclusively for this: automatic encryption, stale connection detection, fast state syncing. It is also the word for a map of neural connections — a set of interconnected nodes you own is an extended nervous system.

  3. 03

    User Engine

    your half

    Expand DMT with your own creativity and desires. Your code lives beside the core rather than inside it, so upgrading the engine never costs you your work. Imagination is the limit.

  4. 04

    DMT API

    the surface

    Take advantage of prepared functionality through dmt/modules and the special program instance. Design once, then free your brain and let the code execute.

The Road Ahead

DMT SYSTEM 1.0 2.0

Where the project actually stands, without the marketing gloss.

2017we are here2.0

v1.2.250

Working Proof of concept

Very limited functionality but it proved the connectivity and interaction model on local networks. That is the honest description: a real system, with real limits, doing real work every day.

v1.5

Unreleased but stable

Almost ten years of use and correction. The mathematical structure was right from the beginning; the rest has been ground down into place one release at a time. Versions 1.5+ are not yet certain to be released (see the letter below).

v2.0

In development

Fixes many known vulnerabilities and greatly expands what the system can do. The entire user experience is on another level. Not yet sure how or when it will be available for general use.

Learn more

When we started around ten years ago it was jokingly said that this would be one of the best projects in the history of personal sovereign computing. And now it actually is.

It is a special kind of project — attention to detail, covering every aspect with the best possible compromises. It worked out with the addition of capable A.I. coding. The mathematical structure was right from the beginning, and it needed a concentrated, multi-month guided push through the most powerful A.I. models to bring the original plan to light with limited human resources.

Open questions

Can new DMT SYSTEM be public?

Open source is strange, and often not properly rewarded. One often drowns in unpaid work where people just demand without contributing more than they take. So in addition to further technical improvements and modernization, this part needs to be figured out if the project is ever going to be available for general [hacker] use.

I will have to decide at some point whether it is worth a try to build a self-sustaining organisation with more people helping with support. More people means income is needed, otherwise the project falls apart. For now I'm incredibly happy where things are, and these are just open questions for later.

Honestly, I'm leaning towards keeping DMT 1.5 to 2.0 private — not because of selfishness, but because properly promoting and explaining the entire system would take a heroic effort. It is more a combination of healthy fear and clear thinking than arrogance or selfishness. For this effort we need more people and more people means having means to pay them for the effort. We don't want pure volunteers because pay is good and makes the modern world possible.

So this is an enormous chicken-and-egg problem, which is also kind of sad if v2.0 indeed does not get more broadly distributed. Once we start in this direction there is no going back, so it is better not to start at the moment. If 10 to 100k right people would properly learn about capabilities of this system, they would together more likely than not gladly provide the neccessary resources for success. For so many people to learn about it we would need the big effort with massive resources, hence circular dependency.

I value the potential benefit for the world in the $42 to $69 x 106 range, and it is mathematically entirely possible to reach and "capture" this. But it's just hard, and I'm not in a big mood for it currently (too busy with development and generally lazy in regard to spreading innovations). When we are at v2.0, this might change. Half of the total is for delivering a DMT SCREEN physical device to each buyer. Meaning half of the worth is hardware and other half is software which runs more broadly on users' personal cloud — DMT NETWORK.

It is truly amazing how it worked out. Hopefully this does not sound too grandiose — there is a thin line between judging people's true happiness and pride in having created something previously uncertain which turned out to just work. The open question now is how (and if) to offer it to the world when it is completely ready.

David

expected distance

The desire for excellence is an essential feature for doing great work. Without such a goal you will tend to wander like a drunken sailor. The sailor takes one step in one direction and the next in some independent direction. As a result the steps tend to cancel each other out, and the expected distance from the starting point is proportional to the square root of the number of steps taken.

With a vision of excellence, and with the goal of doing significant work, there is a tendency for the steps to go in the same direction and thus go a distance proportional to the number of steps taken, which in a lifetime is a large number indeed.