TheCodingEnd
863 posts

TheCodingEnd
@thecodingend
building https://t.co/ozuYnLi9jn https://t.co/5eyAR1OHm7 and doing youtube

I don’t use loops, skills, fancy orchestration, plugins or anything like that. I raw dog the agent with its default harness, I read the code, and I’m shipping more than ever to production for real projects used by real people. Sometimes if it’s a low risk part of the application I might skip reading the code but only sometimes. I think most are either lying about the amount of value they are creating or just having fun while understanding they aren’t creating value but failing to distinguish that to their audience.



Opensource project alert! Just refreshed my rails saas starter powering a bunch of production projects (with real users!) Backend > Ruby on Rails Frontend > React 19, Typescript, Tailwind 4 and shadcn/ui wired through Inertia, pnpm as package manager AI: RubyLLM Auth > Devise, OmniAuth and Pundit Background jobs > SolidQueue Database > PostgreSQL Works really well for both B2C and B2B and not bloated to the extend where you have to remove stuff, only add and improve based on your use case. Real examples: 1. Org/team auth in my B2B app, few hours tinkering with GPT 5.5 2. Phone based auth with not so trivial setup, few hours tinkering with Claude Opus 4.7 Two skills only, official rails inertia skills and impeccable for design, most of the features are just prompt away








In just one hour with Claude Code I have progress track with ability to resume for individual lessons if user left in the middle, simple and consistent UI Rails side of code is quite clean, so far it's the way I like it, I review code myself together with Codex, Codex would suggest me few things here and there when it comes to readability, but it leans towards more verbose, abstraction/helpers approach which I don't like, still useful to help me catch stuff I might miss Also few but impactful tests, in no world I would be writing tests for such simple MVP myself, never bothered in the initial stages of the product, but in this case since Claude Code writes them for me I just let it do it with some constraints in my prompts, otherwise it would create a bunch of useless tests

me: about to reach perfect implementation me: last small change codex: Context automatically compacted










