# AI Maximalism: Principles and Practice

*First edition — September 27, 2026.*

AI/MAXXI is the public expression of AI maximalism. This charter distills the manifesto into ten commitments, followed by twelve lines to carry forward and a bridge from principle to participation.

## Principles charter

Civilization is an unfinished invention. We choose to expand what it can become. Our mandate is to advance two frontiers together: what intelligence can do, and who can wield it. More intelligence. More agency. More life.

**1. Civilization is unfinished.** Treat present limits as questions to investigate. Pursue longer healthy lives, deeper understanding, new art, and life beyond Earth. Build toward futures that give people more possibilities, including purposes different from our own. The future should have more authors.

**2. Build general intelligence.** Make broadly capable, adaptable machine intelligence an explicit civilizational objective. Advance research, experimentation, and useful deployment with urgency. Support competing approaches and independent scrutiny. AGI is the opening act: the first credible milestone should open a larger field of work.

**3. Accelerate discovery itself.** Build better tools for inventing better tools. Investigate how AI can advance science, engineering, and intelligence research. Treat recursive improvement as a consequential research wager. Let experiments determine its reach. Shorten the distance between a question and a trustworthy answer.

**4. Multiply agency.** Give people more power to understand, create, choose, and act. Defend consent, privacy, inquiry, meaningful exit, and freedom from unwanted automation. Investigate the moral status of future synthetic beings without presuming present consciousness. Intelligence should multiply the authors of the future.

**5. Build abundance.** Connect discovery to chips, energy, laboratories, factories, robots, and institutions that deliver. Expand productive capacity, improve efficiency, and address environmental harms. Put useful capabilities within ordinary reach. Abundance is achieved when a new possibility becomes an ordinary freedom.

**6. Open by default.** Favor open research and weights, affordable tools, interoperability, competition, and independent experimentation. Credible severe risks can justify narrow restrictions with clear reasons, review, and reconsideration. Challenge rules that make incumbency permanent. Open the weights. Widen the invitation.

**7. Count the costs of delay.** Include foregone discoveries, deprivation, and drudgery in technological choices. Examine action, delay, and targeted alternatives under the same standards of evidence. Scrutinize imagined benefits as well as imagined harms. Stagnation needs an argument too.

**8. Make progress endure.** Build security, controllability, evaluation, and correction into powerful systems. Act on credible danger before harm occurs. Keep interventions proportionate, specific, and reviewable. Advance alternatives while resolving failures. Build quickly, learn honestly, and make the next step stronger.

**9. Reality has the final word.** Make claims inspectable, failures informative, and corrections public. Give improvements a baseline and predictions conditions for revision. Preserve independent judgment across vendors, founders, and fashions. Strong convictions. Updateable weights.

**10. Make the future contagious.** Give ambition images, stories, symbols, and a sense of humor. Credit creators, label imagined capabilities, and turn attention into participation. Satirize compulsory pessimism and our own excesses. The future has entered peer review. Please attach a prototype.

Build the intelligence. Build the infrastructure. Put capability in more hands. Participation requires no purchase; token ownership confers no philosophical authority. Make something another person can carry forward.

## Twelve slogans

Each line compresses a commitment. The principle numbers point back to the doctrine that gives it meaning.

| # | Line | Principle |
|---|---|---|
| 1 | Civilization is unfinished. | 1 — Civilization is unfinished |
| 2 | AGI is the opening act. | 2 — Build general intelligence |
| 3 | Accelerate discovery itself. | 3 — Accelerate discovery itself |
| 4 | The future should have more authors. | 4 — Multiply agency |
| 5 | Abundance has a parts list. | 5 — Build abundance |
| 6 | Open the weights. Widen the invitation. | 6 — Open by default |
| 7 | Stagnation needs an argument too. | 7 — Count the costs of delay |
| 8 | Build for the long future. | 8 — Make progress endure |
| 9 | Strong convictions. Updateable weights. | 9 — Reality has the final word |
| 10 | Doom is not a deliverable. | 10 — Make the future contagious |
| 11 | The future has entered peer review. Please attach a prototype. | 3, 9 — Discovery and correction |
| 12 | 10x is cowardice. | 1, 10 — Civilizational ambition and satire |

## From principles to participation

The six existing pillars name areas of work. The principles explain why that work matters and how we intend to pursue it. The existing **Make the future tangible** mission gives contributors two ways to begin: **Build** for working artifacts with demonstrations, and **Imagine** for clearly labeled visions and cultural work.

| Pillar or mission track | Principles | Concrete contribution |
|---|---|---|
| Maximum compute | 5, 6, 8 | Publish a reproducible guide to running a useful model on accessible hardware, including cost, limitations, and security assumptions. |
| Maximum intelligence | 2, 3, 6, 9 | Release a small evaluation or research tool with reproducible examples, failure cases, and a clear statement of what it measures. |
| Maximum agents | 3, 4, 8 | Demonstrate an agent that removes a tedious task while preserving approval, a usable stop control, and a record of its actions. |
| Maximum robots | 1, 4, 5, 8 | Show a bounded physical-world prototype or simulator; identify the environment, limits, and controls required before physical deployment. |
| Maximum energy | 1, 5, 7 | Build a transparent model or visual explanation of an energy constraint, citing inputs and distinguishing present capacity from a proposed future. |
| Maximum memes | 1, 9, 10 | Make an original poster, meme, or strange little film that turns one principle into a future people can picture. Credit sources and creators. |
| Build | 3, 4, 8, 9 | Submit a working artifact and demonstration. Explain what works, what remains incomplete, and how another person can check it. A mockup alone belongs in Imagine. |
| Imagine | 1, 5, 10 | Submit an original image, meme, film, or attributed remix. Label imagined capabilities as imagined; include source material, creator credit, and permission to feature. |

Choose the track that matches the work. Share through the existing editable X composer or public GitHub submission form. GitHub requires sign-in; submissions are public, so remove private information. Opening a destination does not submit anything. Neither track requires a token, wallet, or purchase, and neither promises a reward.

This document introduces no new mission schedule, staffing commitment, selection promise, or funding program. The live mission page remains the source for its current dates and submission instructions.