hyper.dev
well come

🖐🏽

2026-11-01 · mezangelle - 1 - Do It Yourself: an artificial intelligence

Written 2017/01/01, before the open letter to Mez Breeze. Republished; the original is archived at the Wayback Machine. Footnotes added in 2026 with Claude.

The main goal of artificial intelligence (AI) is to translate human knowledge into equations.

The problem that arise and still holds is that it can’t be solved on a human time scale with current technology.

The problem must be reduced to a problem that can be solved under time constraints pratical for mortals.

Machine learning produce projections of human knowledge over a smaller number of dimensions to make the computation pratical and sometime infer knowledge about how the results was done.

Instead of trying to put order in the mix, my idea is to increase the entropy of knowledge graph using grounded topologies.1

Building such structure can be done though personnal note taking applications, croudsource applications and puzzle games.

Not only we leverage Human knowledge and computation skills but append to the already mastered ways of the intelligence with improved logical and fuzzy reasoning but also push further the use of pattern matching cognition skills that the brain already has.

The opportunities to create cogni-games2 are manyfolds. People with disabilities can be offered new ways to interact with the world in a meaningful and enjoyable way.

cogni-art and cogni-game are really what should.be+is all about anthroposcene.

Anthropo[morphically_s]c[r]e[e]ne[d]
Anthropoc[Sh]e[e]ne[d]
AnthroP[r]o[bos]c[is]ene
--
AnthropO[bs]cene
--

Anthropo[S]ceney||AnthropO[bs]cene by Mez Breeze

DIY.


  1. grounded entropy (2017: “grounded topologies”, mislabeled) — let the knowledge graph grow as rich and messy as it wants, but let each connection come from a human act: a note, a move in a game, a crowd’s answer. Grounding sounds like it should reduce entropy; that tension is the point. The shape that emerges is the topology.↩︎

  2. cogni-games have prior art: Luis von Ahn’s “games with a purpose” (the ESP Game, 2004, then reCAPTCHA, 2007) and Foldit (2008) already turned play into computation. Those games harvest the player’s work; cogni-games aim for the player to grow too, and open the door to people with disabilities and to art. In 2022 that became: we play, they learn, and we grow together.↩︎