ai.console/ai-atlas/glassbox.trace

AI Atlas

From first principles to production systems.

An executable map of intelligent systems: from vectors, probability, search, objectives, gradients, and learned representations to inference, retrieval, agents, evaluation, safety, and operations.

254Canonical pages
24Branches
4Executable labs
33%ES coverage
GLASSBOX INTELLIGENCE

AI is observable computation, not a magic box.

Each note asks what is represented, computed, optimized, learned, measured, hidden by frameworks, and changed by production constraints.

Learning paths

Choose a route; keep the computation visible.

Every route names prerequisites and crosses theory, implementation, evidence, failure modes, and production behavior.

Open Start Here
SPINE PROJECT · v0→v10

Glassbox AI Lab

Build the stack progressively: scalars and probability → autodiff → models → runtimes → retrieval → agents → production evidence.

Open the lab specification
ATLAS LEGEND

Depth and editorial state

TYPEOVERVIEWFOUNDATIONALDERIVATIONIMPLEMENTATIONSYSTEMPLAYBOOKPAPER GUIDELAB
LEVELBEGINNERINTERMEDIATEADVANCED
STATUSCURRENTREVIEW NEEDEDOUTDATEDPLANNEDEXPERIMENTAL
00
Phase 00 · Map the field

Orientation

How to use the atlas, read its notation, validate claims, and choose a route.

Phase page
01
Phase 01 · Represent and reason

Foundations

Mathematics, computation, classical reasoning, statistical learning, data, assumptions, and uncertainty.

Phase page
02
Phase 02 · Learn representations

Learning & Models

Neural learning, sequential decisions, architectures, language, vision, audio, and multimodality.

Phase page
03
Phase 03 · Train and serve

Training & Inference

Training systems, model adaptation, inference runtimes, hardware, latency, throughput, and cost.

Phase page
04
Phase 04 · Ground and act

Context & Agency

Prompts, assembled context, retrieval, memory, tool use, agents, permissions, and recovery.

Phase page
05
Phase 05 · Measure and trust

Measurement & Trust

Evaluation, interpretability, safety, security, governance, uncertainty, and release evidence.

Phase page
06
Phase 06 · Ship and operate

Product & Operations

Product framing, MLOps, observability, monitoring, feedback, reliability, and human review.

Phase page
Phase ★ · Build and investigate

Labs, Research & Playbooks

Glassbox labs, paper reproduction, research logs, and repeatable playbooks.

Phase page
Reference layer

Registries and operating memory

Registries keep external references, decisions, datasets, tools, and evaluation assets normalized behind short atomic notes.

AI Atlas · Lautaro Damore

A framework-free bilingual atlas by Lautaro Damore. Observable computation, reproducible evidence, production consequences.