IdeaWell

Project ideas from Hacker News discussions.

FTL: A new operating system for clouds

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. Hobby‑vs‑professional perception
- “Is it just a hobby, and won't be big and professional like gnu?” — drybjed
- “Not just a hobbyist, an enthusiast …” — mlinksva

2. Kernel architecture classification
- “It's an exokernel. It's kind of like if you modified a hypervisor specifically for running unikernels instead of classic VMs.” — monocasa
- “So it's a microkernel…ish? And it runs Linux programs and supports enough features to serve its own website. Excellent; I hope it takes off.” — yjftsjthsd-h

3. Language choice and safety considerations
- “The real defensible reason is that with current tools available you can write perfectly safe (safer than Rust even) C++ code, while avoiding the horrid Rust compile times …” — rfgplk
- “Now people just wonder why you didn't write it in rust.” — habitue

🚀 Project Ideas

VeriC: AI‑Assisted Formally Verified C Toolkit for OS/Kernel Development

Summary

  • Provides an AI‑driven pipeline that translates high‑level specifications into memory‑safe C code, automatically generates proofs, fuzz harnesses, and regression tests, targeting OS/kernel developers who want Rust‑level safety without its compile‑time overhead.
  • Core value proposition: dramatically reduces the effort and risk of writing low‑level system code while keeping the performance and familiarity of C.

Details

Key Value
Target Audience OS/kernel engineers, embedded systems developers, hobbyists experimenting with new kernels
Core Feature Spec‑to‑verified‑C conversion with integrated theorem proving, fuzzing, and safety annotations
Tech Stack Rust (for tooling), LLVM/Clang, Z3 or Coq backend, LLM API (e.g., GPT‑4), CI/CD hooks
Difficulty Medium
Monetization Revenue-ready: SaaS subscription (tiered by monthly verification minutes)

Notes

  • HN commenters lamented Rust’s “horrid compile times” and wished for “formally verified C to become the standard” (tkz1312, rfgplk). VeriC directly addresses that desire.
  • By automating proofs and fuzz generation, it lowers the barrier for projects like FTL to achieve memory safety without sacrificing speed, sparking discussion on safety vs. performance trade‑offs.

ExoBuild: Declarative Exokernel/Unikernel Builder with AI‑Generated Drivers

Summary

  • A CLI/web service that lets developers declare the minimal set of kernel capabilities (e.g., networking, storage, security primitives) they need; ExoBuild then strips unnecessary components, synthesizes any missing device drivers via an AI model trained on driver specifications, and outputs a bootable exokernel image capable of running Linux binaries through a thin compatibility layer.
  • Core value proposition: eliminates the tedious driver‑writing and kernel‑configuration steps that plague new OS projects, letting teams focus on application logic.

Details

Key Value
Target Audience Cloud‑native startups, edge‑device firms, OS hobbyists building minimal runtimes
Core Feature Declarative manifest → trimmed exokernel + AI‑synthesized drivers + Linux syscall shim
Tech Stack Go (CLI), Docker for build isolation, LLVM‑based codegen, LLM for driver synthesis, QEMU/Firecracker for testing
Difficulty High
Monetization Revenue-ready: Per‑image build pricing + optional support contracts

Notes

  • Commenters noted the burden of “writing drivers for GPUs, wifi, power/thermal” (mikepurvis) and wished for a way to “run Linux programs with fewer abstractions” (yjftsjthsd‑h). ExoBuild offers exactly that: a path to run Linux workloads on a minimal exokernel without manual driver work.
  • The ability to quote “the diagram renders OKay in Brave” shows interest in clear, shareable outputs; ExoBuild would produce reproducible, verifiable images that could be discussed and benchmarked on HN.

BootApp: Self‑Booting Application Image Generator for Minimal Kernels

Summary

  • Takes an application binary (or source) and a target minimal kernel (e.g., FTL, a custom exokernel) and automatically produces a bootable UEFI/USB image that includes only the needed kernel parts, AI‑generated hardware abstraction layer, and the app, ready to run on bare metal or in a VM with zero OS overhead.
  • Core value proposition: enables developers to ship “game‑on‑a‑USB”‑style appliances or secure single‑purpose appliances without managing a full OS, mirroring the nostalgia of self‑booting PC games while leveraging modern kernel tech.

Details

Key Value
Target Audience Game developers, appliance makers, security‑focused devops, hobbyists
Core Feature App‑specific boot image creation with kernel trimming & AI‑generated HAL
Tech Stack Rust (image assembly), UEFI SDK, QEMU for validation, LLM for HAL generation, GitHub Actions for CI
Difficulty Medium
Monetization Hobby (open‑source core) with optional paid premium templates

Notes

  • The thread revived memories of “self‑booting IBM PC compatible games” (teddyh) and the idea that “every game could come with its own bootable USB” (rfgplk). BootApp makes that vision practical today.
  • By providing a one‑click way to generate secure, minimal boot images, BootApp would spark practical utility discussions (e.g., benchmarks vs. containers) and be loved by commenters excited about “booting straight into a chat prompt”‑style demonstrations.

Read Later