background software discuss links notions changes (printable) (crit me now)
essential glossary goodness how_to whitepaper

How to Build a Global Brain


This is a high level description of a software project currently underway. Do not confuse the casual hubris of this document with groundlessness. There is no profound new AI technology in this effort. No incredible breakthroughs are required. Rather, the essence of this undertaking is the integration of existing technologies carefully selected for the synergistic effects that their union unleashes. Merely an act of emergence by design.

The problem

Existing software is tedious, cumbersome, ineffective and divisive. Documents are dead. Different file formats. Different applications. Different operating systems. Different interfaces. Different places. Dead links. Busy servers. Drafts, revisions, final versions. I left it at home. It wasn't backed up. It doesn't look right on my screen. It needs your input. What is the crux? What do the experts think? What does Bob think? Is there a consensus? Where's the wacky stuff? the new stuff? the good stuff? What constitutes goodness in this field?

Emergence: complexity springs from simplicity

Every complex thing in the universe, and all the trivial ones too, can be seen as the consequence of simpler things just being themselves and their interactions with their peers giving rise to new orders of complexity.

? begat quarks. Quarks begat subatomic particles. Particles begat nuclei. Atoms begat chemistry. Chem begat biochem. Biochem begat biota. Prokaryotes begat eukaryotes. Single celled. Multicelled. Tissues. Organs. Brains. Language. Tools. Writing. Monastaries. Universities. Assembly Lines. Telecommunication. Computers. The Military Industrial Complex. Networks. Internet.

Global Brain.

So grow it, but in what?

One of the clearest successes of the AI community is a technology called knowledge representation. It uses little items called frames which, like tinkertoy hubs, have wee slots on them to connect them to other frames. Most anything that can be said can be expressed as frames. These frames, each representing a single concept or thing, can be connected together and collected together in knowledge bases. These knowledge bases can themselves be linked so that frames can inherit from one another or be grouped together in new ways. Knowledge bases are stored on knowledge servers and can be copied, moved, and shared across servers very reliably.

In short, we can create a container into which people can externalize their knowledge, opinions, perspectives and even feelings. A distributed, self-organizing, load-balancing, lattice of knowledge servers is a suitable substrate for a global brain. Let them mirror and buffer the popular stuff. Let knowledge slosh about the network to where it is wanted. Workgroup servers, personal servers, mirror servers will provide graceful failover when authoritative servers go down. Knowledge bases will be accessed by name rather than visited at addresses, because sometimes individual addresses fail.

Where does the intelligence come from?

Doesn't a "global brain" need some intelligence? Where's that coming from? Artificial selection (not the natural stuff) applied to the ideas (and claims, suggestions, critiques, versions and new features) contained in Nooron. Let good ideas inspire better ones. Let bad ones pass unnoticed or be gone over postmortem. Human minds are great at discriminating, dandy at comparing, eager to share their opinions, endlessly innovative, easily inspired by others. In short, there might be enough intelligence in the world to solve some of the big problems, to create some stupendous works. All we need is some software which can get out of the way and let those minds work together on both the big picture and the details.

Q. What will we see? How will we use it?

Live charts, graphs, networks, flowcharts, diagrams and automatically formatted text will flow naturally out of the knowledge. The content of these views will be easily tailored by each of us via filters and mappings to depict particular perspectives on the knowledge. These depictions will be saveable, shareable and subject to evolution -- just as the ideas are -- by being enjoyed, recommended and improved by others. We'll even be able to evolve world views: collections of filters and lenses that can be used to view different knowledge bases; to bring a familiar perspective to bear on new data or fresh perspectives on the familiar. We'll find like minds. We'll find foreign ones. We'll make friends and influence people. We will express values, clarify objectives, discover consensus, organize efforts and get things done. We will accelerate the evolution of ideas.

Q. How will it feel? (to us)

  • It will feel like playing with an intelligent graphical toy.
  • It will feel like participating in a collective intelligence.
  • It will feel like a barn raising. In the sky.

Q. How will it feel? A. Through us.

What will it feel? Will there be an it 1, like which it will be something to be? Will it just be the data in the system? or will it be the sum of the software, the data and its users?

Status: circa January 2003

See /software for a history. A useful (but incomplete) implementation is available at

  1. What is it like to be a bat? by Thomas Nagel
Shawn Murphy
Created: Wed Jun 3 09:57:48 MDT 1998
Last modified: 2003/01/31 01:36:52.6987 US/Mountain