// journal · field notes
Field notes from a solo builder. AI, product, engineering, and the unglamorous parts that actually move the needle.
// all articles

Copy-ready prompts for migrating a Lovable.dev export to the Next.js App Router in safe slices: auth, billing, SEO pages, dashboards, checks, and rollback.

A practical decision and sequencing guide for converting a Lovable React app to the Next.js App Router without breaking auth, SEO, billing, or launch redirects.

Vibe coding is a real way to build software by letting AI generate most of the code while you steer through prompts, feedback, and taste. Here is what it actually means.

The best vibe coding tools in 2026 organized by project stage: UI prototypes, existing codebases, planning, zero-friction experiments, and deployment.

Cursor vs Windsurf is not a permanent identity choice. The honest breakdown: Cursor for existing codebases, Windsurf for greenfield builds. Here is why.

Cursor IDE is an AI-native code editor built on VS Code. Here is when it genuinely helps, where it falls short, and whether it belongs in your daily workflow.

An engineer who ships production AI apps explains what the best AI app builder means in 2026 -- decision matrix, failure modes, and real pricing.
Decision framework for choosing between a Lovable agency and a freelance Lovable expert, with honest rate comparisons and failure modes.
AI pair programming has matured past autocomplete. A practitioner's guide to the four modes, the tools that actually work, the failure modes nobody talks about, and how to keep humans and bots on the same contract.

AI developer rates in 2026: £80–£400/hr depending on seniority, geography, and engagement shape. Tables, hidden costs, and honest comparisons.
Lovable, Bolt, and Cursor compared by someone who has shipped production apps with all three. Honest trade-offs, exact failure modes, and a decision matrix.

An engineer shipping with Cursor, Claude Code, Copilot, Windsurf, Aider, and Continue ranks them honestly -- real pricing, comparison table, decision tree.
The playbook we use when founders call us with a broken Lovable app in production: the 6 ways they break, a 30-minute audit, and the fix patterns.
An engineer's honest Lovable review after 20+ production projects. What it nails, where it breaks, real 2026 pricing, and who should and shouldn't use it.
Honest, phase-by-phase guide to migrating a Lovable SPA to Next.js — covering SSR, Supabase SSR auth, SEO, and real time estimates by app size.

A complete decision framework for hiring an AI developer in 2026 — engagement models, interview questions, red flags, portfolio review, and rate expectations.
Lovable builds React SPAs that Googlebot sees as an empty div. Three real ways to fix it: edge worker prerender, Next.js migration, or Prerender.io.
Seven honest Lovable alternatives compared by a practitioner who ships with all of them. Covers Bolt, Cursor, v0, Replit, Base44, and when to use none of the above.
A practitioner's guide to hiring a Lovable developer: skills that matter, where to find them, rates, and red flags.

A field report on agentic engineering: 99 PRs, 230 commits, 80,000 LoC across 12 production repos in 24 hours — and the seven orchestration patterns that make it possible.
Stage-by-stage staffing map: pre-seed freelance, seed founding engineer, Series A platform team — 2026 rates, skills by stage, and the token-cost math that kills AI products.
The filter I use: scoping, skills checklist, sourcing, 5 interview questions with real answers, 2026 rate bands, and the 3 red flags that have burned every client.
The actual working config for ~/.cursor/mcp.json — two transports, env-var secrets, the 5-minute setup flow, and the debugging loop for the red dot.
MCP is not a REST replacement. Decision table, real code from both, side-by-side comparison across 10 dimensions, and the hybrid pattern that ships in production.
Not a list of cool MCPs to try — the 12 I actually run in Cursor every day, with the real mcp.json config for each and notes on which transport to pick.
For 80% of AI projects under $150k, a vetted freelance senior ships faster and cheaper. Honest 2026 rates, the hybrid model that beats both, and a decision checklist.
Four layers every production MCP server needs — transport, JSON-RPC shape, tool definitions, and guardrails. With real code for rate limiting, circuit breakers, and structured errors.
Six lines of instructions, zero information loss, and nearly 10% off my AI bill. The exact rule, the numbers behind it, and why verbose responses bleed budget twice.
Everyone's worried about being replaced by machines. The real threat is complacency. AI is a power tool — dangerous in lazy hands, unstoppable in skilled ones.
Discover how we transformed our VS Code experience with custom extensions tailored to our workflow frustrations.
Transform repetitive development tasks into reusable, one-command automations using Cursor skills, saving time and ensuring consistency across your projects.
Automate your pre-push workflow with Ship It for seamless code quality and consistency. Experience faster, safer deployments in JavaScript/TypeScript projects.
Discover how CaptureContext bridges the gap in AI front-end development by providing AI bots with essential visual context, improving bug diagnosis and resolution.
Exploring how we built a production-stable real-time sync between a web app and a desktop app using WebSocket. Key takeaways include dual-channel architecture, handling reconnection, and ensuring stability.
Our journey from scattered secrets across various environments to a unified Infisical setup, with AI-driven migrations and MCPs.
Struggling with scattered MCP configs, agent modes, and tool workflows, I centralized everything into a versioned foundation repository with auto-updates and visibility dashboards. This transformed a fragile, six-location mess into a predictable, auto-synced system that cuts setup time from 45 minutes to zero and surfaces real tool health instead of silent failures.
Practical techniques for achieving perfect Lighthouse scores in production. No tricks, just engineering.
Learn how to design, test, and deploy reliable multi-agent systems that scale. Practical patterns from real production deployments.
How to build rock-solid applications with Fly.io and Supabase. Architecture patterns, deployment strategies, and common pitfalls.
Good design isn't about making things pretty. It's about making hard choices — what to cut, what to keep, and what to make impossible to ignore.
No team. No meetings. No excuses. The exact toolkit — from IDE to deploy pipeline — that lets one person ship production apps in under a week.
Developers who pitch architecture lose deals. Developers who pitch outcomes win them. How to reframe your work in the metrics clients actually care about.
The graveyard of great ideas is full of perfect plans that never launched. Why velocity beats polish, and how to rewire your brain for output over overthinking.
The four boring changes — parallel tests, Docker layer caching, Terraform, live metrics — that cut our CI/CD pipeline from over an hour to six minutes.
// subscribe
One email when there's something worth reading. Unsubscribe in one click.