Code Execution & Runtimes

The hardest invisible problem in the stack

The runtime is the foundational platform decision. Master the tradeoffs between server-side containers, browser-based execution, serverless and edge runtimes, and sandboxing mechanisms.

Modules in This Level

2.1 What a Runtime Actually Is

Understand the three runtime layers — engine, standard library, and runtime APIs — and why the differences between Node.js, Deno, Bun, and browser environments determine which npm packages work, what system calls are permitted, and what Make's cold start time will be.

2.2 Server-Side Containers (Docker)

Learn how Linux namespaces, cgroups, and layered filesystems combine to make Docker containers the standard execution unit for cloud workloads — and when containers are the wrong answer for Make's preview infrastructure because they share the host kernel.

2.3 Browser-Based Execution (WebContainers)

Learn how WebContainers runs Node.js inside the browser via WebAssembly — intercepting POSIX system calls, emulating a filesystem, routing network through a ServiceWorker — and what native module incompatibility means for Make's package support decisions.

2.4 Serverless Functions & Edge Runtimes

Compare AWS Lambda's Firecracker isolation, Vercel Functions' Next.js-aware layer, and Cloudflare Workers' V8-isolate model — and understand why Make's choice of API route execution environment constrains which npm packages, runtime APIs, and cold-start budgets are available.

2.5 Sandboxing and Isolation

Build Make's threat model for untrusted AI-generated code and evaluate the five-level isolation spectrum — from iframe sandbox to Firecracker microVM — to understand the startup time, cost, and security tradeoffs that determine the right containment strategy per workload type.

2.6 Cold Starts, Warm Pools, and Latency Budgets

Decompose the full latency pipeline from agent inference through dependency install, runtime startup, HMR update, and browser repaint — and learn to identify which stage consumes the most of Make's 2-second preview budget so platform investment targets the right bottleneck.

2.7 Framework Support as a Platform Decision

Learn what "supporting a framework" requires at the platform level — templates, dependency caches, dev server config, build pipelines, deployment adapters, and LLM training data — and why adding Vue or Svelte multiplies operational complexity faster than demand justifies.

2.8 Level 2 Checkpoint

Apply Level 2 runtime knowledge to an engineering design review: evaluate a proposal to migrate Make's preview infrastructure from Docker to browser-based WebContainers, weighing cost savings against native module compatibility, vendor dependency, and security tradeoffs.