// the manifesto

What we are,
and what we refuse to become.

no-magic-ai is an umbrella for studying how modern AI actually works. Not a framework. Not a course platform. Not a chatbot. Every repo under this org enforces a single-sentence constraint. Every algorithm points back to a paper. Everything else is off-charter.

// mission

Because model.fit() isn't an explanation.

Modern AI is taught as a stack of frameworks. transformers.AutoModel. trl.PPOTrainer. peft.LoraConfig. Contributors learn to call the library, not to derive the algorithm. no-magic-ai exists to close that loop — by implementing every algorithm from its source paper, in Python standard library only, one file at a time, so a determined reader can trace every tensor back to its equation.

Our audience is two people: the ML engineer who uses frameworks daily and wants the internals, and the career switcher who read the papers and needs a working implementation they can run on a laptop. Everything we build serves one of these two. When it stops serving them, we delete it.

// the core discipline

One constraint per repo. No bleed.

no-magic-ai expands through new repos, not feature sprawl. Every repo declares one sentence that governs what belongs and what doesn't. Repos do not share build systems, do not import each other, do not cross-reference code. The website is the only surface that joins them.

no-magic
One algorithm per file. Stdlib only. CPU only. Runs in under 10 minutes.
live
no-magic-viz
One Manim scene per algorithm. Renders to MP4 and GIF.
live
no-magic-ai.github.io
Static HTML portal. No JS framework. Federates the ecosystem.
live
no-magic-papers
One markdown per paper, plus companion lessons. Summary, contribution, status, link to implementation.
live
no-magic-paths
Curated reading orders with prerequisites across algorithms and lessons.
planned
no-magic-labs
Exercises and applied projects that exceed one-file scope.
later
apprentice
Agent pipeline that drafts algorithm entries from papers for human review.
design

// the hard no's

What no-magic-ai refuses to become.

The value of this org is in what it won't do. Every feature we reject keeps the scope tight enough that the code stays readable and the constraints stay honest. These are not deferred — they are off-charter.

not an LMS

No user accounts. No progress tracking. No auth. No server-side state. Static files and GitHub Discussions are the whole platform.

not a chatbot

No RAG over the catalog. No AI tutor. A good single-file implementation with the right comments is already the explanation.

not a framework

No shared base classes. No abstract interfaces. No no_magic.core package. Every script stands alone.

not a tutorial mirror

No translations or ports of fast.ai, Karpathy, HuggingFace, or dive-into-llms. We implement from primary sources only.

not a production library

The code is pedagogical. It is not performant, not distributed, not hardened, and not meant to be imported by your app.

not a benchmark

We show how algorithms work, not which beats which. We cite benchmark results from papers; we do not run leaderboards.

// org-wide principles

The constraints that hold across every repo.

Paper-first sourcing

Every implementation cites the original paper. No content is ported from tutorials, courses, or other educational repos.

Single-sentence charter per repo

A new repo requires a one-sentence constraint that governs it. Violating the constraint is grounds for PR rejection.

Slug as the common key

Each algorithm has a canonical slug (e.g. microrope). Every repo references that slug. No other coupling exists.

Reproducible under commodity hardware

Everything runs on a laptop CPU. No GPU requirement. No cloud credits. No dataset larger than a few MB.

Comments are the curriculum

Code is the primary teaching artifact. Prose explanations live in no-magic-papers/lessons/ and remain optional, never authoritative.

Provenance over abstraction

We'd rather write three similar implementations than one abstracted parent class. Duplication is honest; abstraction hides.

Deletion is a feature

Scripts, lessons, and paths that stop serving the two audiences get removed. The catalog is curated, not accreted.

Static or nothing

No databases. No services. No background jobs. If a feature needs stateful infrastructure, it does not belong in this org.

// sourcing policy

Where our content comes from.

The most important operating rule of this org is what we read while we write. Implementing from the wrong source creates legal exposure and pedagogical noise. These five rules are binding on every contributor and every agent that operates on these repos.

  1. Paper-first. Every algorithm implementation cites the original paper in its thesis docstring. If no paper exists, the algorithm does not belong in the catalog.
  2. No tutorial contact. Do not read other educational repos (fast.ai notebooks, Karpathy's nn-zero-to-hero, HuggingFace courses, dive-into-llms) while writing a no-magic script. Topic inspiration from their tables of contents is permitted.
  3. Attribution is a line, not a link. Cite sources in docstrings with full bibliographic detail. Avoid opaque URL references that can rot.
  4. License vigilance. Before adopting any external asset (dataset, image, figure, text fragment), verify and record the license in ASSETS.md.
  5. Contributor pledge. CONTRIBUTING.md requires every contributor to affirm they implemented from papers, not from tutorials. Enforcement is by review, not by CI.
// on dive-into-llms specifically

The dive-into-llms repository (SJTU BCMI) is an excellent curriculum signal but has no LICENSE file. Its code and prose cannot be redistributed. no-magic-ai has chosen to use its table of contents as a topic menu while implementing strictly from the primary papers that each chapter cites. No code, prose, or translated content enters this org.

// governance

How decisions get made.

The project is small enough today that decisions are captured in documents rather than processes. Documents are versioned, readable by agents, and form the standing record. A decision is binding once it lands in the document; disagreement is handled by opening an issue and amending the document.

Strategy

no-magic-ai-expansion-strategy.md is the top-level doc. Every new repo or charter change references it.

Paper ingestion

paper-ingestion-process.md describes how a new paper flows through the ecosystem from publication to implementation.

Per-repo CONTRIBUTING

Each repo owns a CONTRIBUTING.md that encodes its single-sentence constraint and review checklist.

GitHub Discussions

Open questions, proposals, and curriculum debates live in no-magic Discussions. Resolved items migrate into documents.

// participate

Getting involved.

Contributors are welcome. Implementations, lessons, paper summaries, and curriculum paths are all valid contribution surfaces — each gated by its repo's single-sentence constraint. Read the strategy document, pick a candidate, open an issue before writing code.