Skip to content

Learning Atlas

28 learning paths · 186 ordered steps · 365 glossary terms. Pick a path on the left, or click any term on the map.

This is the Learning Atlas from Orbit. Get your invite

A learning path is a short, ordered run through the glossary. Each step says which term comes next and why it comes next, so a topic you have never touched has a starting point rather than an alphabet.

  • First local-to-remote Git sync

    New to Git? Follow one change from your working directory through staging, a commit and a push, and learn how a local repository syncs with a remote.

    6 steps Git fundamentals
  • Branch work into review

    For developers starting to collaborate: learn how branches, merges, conflicts and pull requests turn solo work into reviewed changes on GitHub.

    5 steps Git fundamentalsVersion control and collaboration
  • Thinking effort: quick answers or careful ones

    For anyone choosing an AI model: learn when a fast chat model is enough, when a reasoning model earns its cost, and how effort and token budgets fit in.

    5 steps AI providers and models
  • One model ladder: Claude and GPT

    Compare the Claude and GPT model families rung by rung, from Haiku and Luna up to Fable and Astra, and see where Claude Code and Codex CLI fit on top.

    13 steps AI providers and modelsAI agents and harnesses
  • Retrieval memory for long-context work

    For builders giving AI long-term memory: learn how embeddings, RAG and vector databases like Pinecone and Chroma work around a model's context window.

    6 steps AI providers and modelsHostinginfrastructuredatabases
  • How Orbit agents use tools

    See how an Orbit agent works: Apollo, MCP, tool use and function calling, and the agentic loop that keeps a coding agent going within a token budget.

    6 steps Orbit-internal conceptsAI agents and harnessesConcepts and methodology
  • Coding harness landscape

    A tour of AI coding agents for developers picking one: Codex CLI, Claude Code, Aider, Cline, Cursor, Windsurf and Devin, and how each one works.

    7 steps AI agents and harnesses
  • From app prompt to hosted UI

    For vibe coders: go from an app prompt in Bolt, Lovable or v0 to shadcn/ui and Tailwind components, a Next.js app, and a live deployment on Vercel.

    7 steps AI agents and harnessesBrowsersruntimesframeworksBuild/deploy tooling
  • Running Your First Local Model

    Run your first AI model on your own computer: learn Ollama, llama.cpp, GGUF files and quantization, plus when vLLM and Hugging Face come into play.

    6 steps Local models and inferenceHostinginfrastructuredatabases
  • Mobile app release tracks

    For app developers shipping to phones: learn Capacitor, React Native and Expo, then automate releases with Fastlane to TestFlight and Play Store testing.

    7 steps Build/deploy toolingBrowsersruntimesframeworks
  • Frontend framework to edge hosting

    For web developers: follow a React and Next.js frontend onto edge hosting, compare Vercel, Netlify and Cloudflare Pages, and see how DNS ties it together.

    6 steps BrowsersruntimesframeworksBuild/deploy toolingMisc tech terms
  • Containers to backend deployment

    For backend developers: learn Linux, Docker, Dockerfiles and Compose, then compare running containers on Kubernetes with hosted Railway and Render.

    7 steps Operating systems and shellsBuild/deploy toolingHostinginfrastructuredatabases
  • Image generation concepts and tools

    For creators new to AI images: learn how DALL-E, FLUX, Stable Diffusion and ComfyUI work, and what diffusion models, latent space and GANs actually are.

    7 steps Imagevideoaudio AIAI providers and models
  • Design canvas to web prototype

    For designers moving into the web: go from a Figma or Spline canvas to Framer and Webflow sites, then Bubble and Retool apps with real logic behind them.

    6 steps Design and creative toolsNo-code / low-code platforms
  • Automating internal operations

    For teams automating internal work: learn how Retool, Make, Zapier, Airtable, Notion and Coda connect into internal tools and no-code workflows.

    6 steps No-code / low-code platforms
  • No-code app frontends and hosted data

    Build a mobile app without much code: compare Glide, Adalo and FlutterFlow, then learn the hosted backends behind them, from Firebase and Supabase on.

    7 steps No-code / low-code platformsHostinginfrastructuredatabasesBrowsersruntimesframeworks
  • Notebook experiments to model assets

    For data and ML learners: move from Jupyter and Google Colab experiments to Hugging Face, then pack a model as a GGUF file and run it in LM Studio.

    5 steps Notebook environmentsHostinginfrastructuredatabasesLocal models and inference
  • Notebook results into shareable docs

    For analysts sharing results: turn notebook work in Jupyter, marimo or Observable into readable documents with Markdown, MDX and React components.

    6 steps Notebook environmentsMisc tech termsBrowsersruntimesframeworks
  • Planning and collaboration tools around code

    A map of the tools around the code for small teams: GitLab and Bitbucket, Jira, Linear, Asana and Trello for planning, and Slack and Discord for talk.

    8 steps Version control and collaboration
  • Automated quality signals before shipping

    Learn the automated checks that catch bugs before shipping: QA scripts, static analysis, linting and type-checking, wired into CI/CD and GitHub Actions.

    7 steps Orbit-internal conceptsConcepts and methodologyBuild/deploy tooling
  • Moving between Windows and Unix shells

    For developers switching operating systems: learn Windows, PowerShell and Bash, run Linux inside Windows with WSL, and find your way around macOS and Zsh.

    7 steps Operating systems and shells
  • API traffic and authentication basics

    For developers calling their first APIs: learn REST, HTTP and HTTPS, why CORS blocks requests, and how OAuth, JWTs and sessions keep users signed in.

    7 steps Misc tech termsConcepts and methodology
  • From secret to least privilege

    Learn to handle secrets safely: API keys and environment variables, .gitignore and Git history leaks, revoking and rotating keys, and least privilege.

    9 steps Misc tech termsConcepts and methodologyGit fundamentals
  • The right way to write a prompt: talk, loop, or spell it out

    For anyone working with coding agents: learn when to chat, when to let an agent loop, and when to write a full spec with failure conditions and a done check.

    8 steps AI agents and harnessesConcepts and methodologyOrbit-internal concepts
  • When your app gets popular

    For new app builders: learn caches, rate limits, load balancers, reverse proxies and API gateways as your app starts handling more visitors.

    5 steps HostinginfrastructuredatabasesConcepts and methodology
  • When a call fails: timeouts, retries and doing it once

    Learn how timeouts, retries, backoff, circuit breakers, idempotency and webhooks help a checkout handle failures without charging someone twice.

    6 steps Concepts and methodology
  • Do it later: queues and events

    For new app builders: move slow work out of signup and checkout with background jobs, message queues, publish/subscribe messaging, events and webhooks.

    5 steps HostinginfrastructuredatabasesConcepts and methodology
  • One app or many services?

    Compare one deployable app with separate services, then learn service discovery, Kubernetes, distributed transactions and saga compensation.

    6 steps Concepts and methodologyBuild/deploy toolingHostinginfrastructuredatabases