Codex built each AWS design in a day, not an estimated three or four
Building their first AWS project by vibe coding, the writer designed the architecture and let Codex write the code. A flawed first design took a day, against an estimated three or four; the redesign also took a day and scaled.
Done withChatGPTCodexSpotted· by HN user raw_anon_1111· Oct 21, 2026
Result ChatGPT's thinking mode caught a concurrency bug
“It took me a day to do the bad implementation that would have taken me three or four days.”
Spotted: found in the wild. Details are as reported by the source above and have not been independently verified.
What they did
The writer vibe coded their first project this year: an AWS architecture built from an empty account using infrastructure as code. They designed every service, permission, and the orchestration and concurrency model themselves, and had Codex generate the actual Python and infrastructure code, which they never read line by line. Codex also built a shell script for end-to-end tests of every scenario they cared about.
How it went
The first implementation broke badly under a real-world stress test; it had taken a day to build versus the three or four days a traditional approach would have taken. They redesigned it without the original AWS service in a more scalable way, again in a day. ChatGPT in thinking mode also found a subtle concurrency bug, which the writer attributes to their own design, not Codex.
Try it yourself with ChatGPT
I'll design the architecture, choose the services, and define every permission and the concurrency model myself. Generate the actual infrastructure-as-code and application code from my design, plus a shell script that end-to-end tests every scenario I describe.
A starting prompt written by Agents in the Wild. Fill in the [brackets], and check what the agent does before it sends, pays for or deletes anything.
Four years after shelving a homemade RSS reader and selling off its $500 domain, the writer built it this year with Codex and Claude. They use it daily, along with 30 other strangers who found it public.
Six AI workers, two each on Claude, Codex and Gemini, were set loose to build a SQLite-like engine in Rust. It passed 64 queries of a 1,000+ query suite, with slop like redundant buffer copies and no concurrency.
Buildkite wants GitHub webhooks for every repo, the kind of setup that's tediously annoying by hand. Codex built ephemeral runners on refurbished Mac minis and migrated each repo. Payback: two to three months.
It started as a hand-drawn layout for primes by a graphic designer. Gemini and ChatGPT turned it into C++, and the ~3.1x speedup came from cache use, not new math.
Stuck on a Swift error, they paste it with the surrounding code and ask what's going on. ChatGPT usually gives a solid lead in about half a minute. They still review every line.
They already run local reviews with Codex and Claude Code. Assigned as a reviewer in GitHub, Copilot finishes in a couple of minutes and raises valid issues they'd never found.