Everything you need to cover EducateWright — with the seams showing.

Download the kit

The full kit, four languages, one click.

The complete media kit — announcement, 101, book draft, provenance, and the True / Honest / Meta ledger — as a print-ready PDF in each language. Or take everything (all PDFs, all logos, the book draft, and the sources) in one archive.

At a glance

Who

EducateWright, a pre-formation nonprofit initiative

Founding partners

SolutionWright Universal and Universal Natural Intelligence (UNI)

Funding commitment

$25 million pledged by founding partners

Allocation

$20M educational grants & scholarships · $5M fully-subsidized mentorships

UNI Workshop

$75,000; partial-to-full-ride scholarships

Foundational paper

Polzin et al., Zenodo DOI 10.5281/zenodo.19785799 (MIT license, 26 April 2026); preprint, Layer 1 complete, Layer 2 review pending

Canonical release

PRWeb, July 9, 2026 (SOURCE: SolutionWright Universal)

A plain-language 101

UNI & active inference, in five steps.

UNI is an architecture built from standard active inference and the Free Energy Principle — the established science associated with Karl Friston and colleagues. An agent does not answer a prompt; it runs a loop.

1. 01

   Perceive

   observation
   Take in what actually happened, not what it hoped would happen.
2. 02

   Infer

   belief update
   Update beliefs to reduce variational free energy — the gap between prediction and what was sensed.
3. 03

   Evaluate

   policy selection
   Score policies by expected free energy: progress toward the goal against reducing uncertainty.
4. 04

   Act

   action
   Commit to the policy expected to reduce surprise, producing the next observation.
5. 05

   Learn

   model update
   Revise the generative model and priors, so the next pass is better calibrated.

The idea is old and cited: from Ibn al-Haytham (c. 1040) and Helmholtz (1867) through Rao & Ballard (1999) to Friston (2010). A fuller history is at universalnaturalintelligence.com/history. What UNI is not: a new field, a new formalism, or a claim of general intelligence, consciousness, or biological equivalence.

Grow bigger worlds

A book, and an invitation to artists.

EducateWright’s story rests on Michael Polzin’s children’s book Big Brains Need Big Worlds (EN/HI/ES, read it here). Its moral is the whole thesis in one line:

“We notice when a child is shaped wrong for their cage — and we change the cage. Not the child.”

The invitation (draft). We are inviting artists — of any medium — to sponsor a child: to teach a form of expression and help what the child’s natural curiosity makes become shareable. An illustration, a song, a poem, a sculpture, a digital work or NFT — any expression at all. The child creates and keeps the work; the artist mentors; EducateWright funds the seat, safeguards the child, and gathers the works into a growing library of natural curiosity. This is a draft invitation, consistent with our pre-formation status; safeguards and terms will be confirmed as the initiative forms.

The claims ledger

True, Honest, Meta — sorted, and explained.

So a reader can tell at a glance what is checkable now, what is a good-faith commitment not yet independently verified, and what is framing about the work and its boundaries.

TRUE — checkable now

The paper

Exists, titled as quoted, deposited 26 April 2026, MIT license. Check: resolve the DOI.

The release

Distributed via PRWeb, 9 July 2026, SOURCE: SolutionWright Universal. Check: open it.

The science

Active inference / the Free Energy Principle is established, associated with Friston. Check: Friston 2010.

The labs

Run in the browser and collect nothing. Check: open them and view source.

HONEST — good-faith, not yet independently verified

$25M pledged

A commitment offered outward by the founding partners — not audited, deployed funds in a formed entity.

Applications open

An open intake, offered in good faith; outcomes depend on review.

Reproducibility

87 pytest assertions across 11 demonstrations on three OSes, as reported; Layer 1 reported complete, not independently re-run here.

Partner affiliations

Quotes reproduced as attributed in the release; company affiliations are not independently corroborated here.

META — framing, provenance & boundaries

Preprint

The paper is a preprint (v0.1.0), not peer-reviewed; Layer 2 human review is pending.

AI co-authors

The paper's co-authors include AI systems (Claude; two OpenAI Custom GPTs), not human collaborators — an organic + AI operator collaboration, as the title says.

Outer edge

UNI is not a new field, a new formalism, or a claim of general intelligence, consciousness, or biological equivalence.

Pre-formation

Entity status, governance, and any named partners will be confirmed after factual and legal review.

No outcome claims

No claim that a tool made a neighborhood safer, that a benchmark is won, or that any outcome is finished.

Research, video & provenance

Logos

The marks, and the family.

Media contact

This kit, like everything we publish, is meant to be checked. If a line overstates, it gets corrected in the open. Grow the world. Grow it together. No one left outside.