Three months of an open source repo built mostly with coding agents. Most of my own pull requests merged without a teammate's approval, so I looked at what actually verifies the code: lint rules that started as review comments, database constraints, agent QA against a real database, and LLM reviewers that are never a required check.
Blog / Dyvertex
Notatki z budowania oprogramowania.
Wyceniliśmy 13 dostawców hostingu, porównaliśmy Prisma, TypeORM i Drizzle, opisaliśmy agentów AI do kodu. Do tego historia, jak jeden git push --all przy GitHub Actions odpalanych na push prawie zrujnował klienta. Wpisy są po angielsku.
I replayed a month of my Claude Code logs. Most of the cost is the model re-reading context, so one setting that compacts earlier cut the cost of a request by 39%. The first three mods I wrote: under 1%.
Running Claude Code and Codex in parallel worktrees moved the bottleneck from writing code to reading it. So I built Treeix, a free, open source AI workspace for Mac: every agent's diff, pull request and ticket in one app, and my comments go back as the next prompt.
One real app - 5,000 users, 1.6M requests, 108 GB of egress - priced on Hetzner, Railway, Render, Fly, Vercel, Cloudflare, Supabase, Firebase, DigitalOcean, Heroku, Sevalla, AWS and GCP with August 2026 list prices. Then priced again at ten times the traffic, because the slope matters more than the starting point.
Different teammates use different AI agents, and each reads its own instruction file. rulesync keeps one set of rules, subagents, commands, and MCP config in sync across all of them.
I ran 12 scenarios on Railway against all three ORMs, dug into pg_stat_statements, fixed the unfair bits, and learned exactly why Prisma is slower, why TypeORM is leaner, and where Drizzle actually beats both.
After using Deno at work, here's my honest take on whether I'd pick it over Node.js for a personal project - spoiler: it's complicated.
Recently, I decided to bring one of my countless startup ideas to life, and I got stuck on choosing a communication protocol between the mobile app and the backend...
Every project begins with a simple idea and limited resources. Our startup was no exception...
Long story short, I almost bankrupted our client...
The short answer is no. The long one - it depends...
Bezpłatna konsultacja · 20 minut
Masz podobny problem we własnej aplikacji?
Umów się na darmową 20-minutową rozmowę audytową i wyślij repo albo błąd. Wyjdziesz z wiedzą, co naprawić najpierw, niezależnie od tego, czy z nami zaczniesz pracę.
Najszybciej odpiszemy na Telegramie: @dyvertex albo mailem [email protected]