[EPIC] Use-case hub + success stories

[stubbi]

Context

We communicate use cases only as a small list on the homepage today. This epic builds the primary growth surface: a /use-cases hub plus per-job pages named by the job (the Daytona and Replit pattern), backed by runnable examples in the engine repo, plus the first case study and the flagship engineering post. Each page names the job, shows the one-line hook with the latency number embedded in the code, proves isolation, and links a runnable example.

The strategic spine (from the use-case and positioning research)

• Category narrative everyone now uses: "a computer per agent." The one validated way to explain fast forking is the fork(2) system call applied to a whole computer: subprocess speed with full hypervisor isolation. Lead with it.
• Closest competitor is Daytona (fork + snapshot + per-agent-computer). Our wedge: microVM by default (Daytona's default is a container; isolation depth ranks Firecracker microVM > gVisor > container), live fork-to-many shipping today (Daytona's parallel-fork API is still partly roadmap), and OSS + self-host + Kubernetes-native (Modal cannot self-host and frames roll-your-own-k8s as a steep ops burden, which is exactly the burden we remove).
• Honesty rule: competitor latency figures are their published numbers on different hardware; ours stay reproducible from bench/. Never describe a system that does not exist.

Sub-issues

• /use-cases hub + data-driven per-job pages #40 /use-cases hub + per-job pages
• Use-case page: drop-in sandbox for coding agents (Claude Code + opencode) #41 coding-agent drop-in (Claude Code + opencode)
• Use-case page: self-hosted cloud dev environments (Coder / Daytona alternative) #42 self-hosted cloud dev environments
• Use-case page: parallel agent fork / best-of-N / tree-of-thought (live-fork flagship) #43 parallel agent fork / best-of-N (flagship)
• Use-case page: RL environments and agent evals (SWE-bench) #44 RL environments and evals
• Use-case page: background agents (task in, PR out) #45 background agents (task to PR)
• Use-case page: untrusted code execution as a security primitive #46 untrusted code execution security primitive
• First case study in the 01 Challenge / 02 Solution / 03 Result metric format #47 first case study (01/02/03 format)
• Engineering blog: How we fork a running microVM in 27ms #48 engineering blog: how we fork in 27ms

Backed by engine work

• fork(n) ergonomics: one call returns N live children mitos#311 (fork ergonomics), #312 (workspace claim), #313 (examples), #315 (benchmarks).

Related existing work (do not duplicate)

• Add real social proof: usage numbers + one named testimonial #15 add real social proof. Engineering blog + human changelog as GTM (flagship "how the 27ms fork works" post) #17 engineering blog as GTM. Comparison pages (mitos vs E2B / Modal / Daytona) #20 / Iterate on the alternatives + comparison pages (/alternatives, /compare/*) #23 compare and alternatives pages. Messaging: make security/isolation a first-class pillar (without losing the fork/swarm lead) #27 security as a pillar.