112 posts · Page 3/4
(*) — Journal

Journal.

Notes on web, automation and AI from our team.

(67) Attio vs Folk — modern CRMs reviewed Two new-generation CRMs that target small modern teams. Polished, opinionated, very different. → (68) App Store Connect changes 2026 Apple updated developer tools. New TestFlight features, AI-powered review insights, alternative store distribution flow. → (69) SEO after AI Overviews — what changed Google's AI Overviews ate top-of-funnel queries. Click-through rates on informational searches dropped 30-60%. Where SEO still wins. → (70) Monday.com as CRM — the real experience Monday started as project management. They built a CRM. Some teams use it. Most regret it within a year. → (71) Hero section patterns that work in 2026 Five hero patterns that landed across 50+ sites we shipped. When each one works, when it kills conversion. → (72) CRM import mistakes that cost real money Importing 50,000 contacts looks simple. Here are 7 ways teams break their CRM in the first hour. → (73) Designing for retention vs acquisition Acquisition design wants to dazzle. Retention design wants to disappear. Most products optimise the wrong one. → (74) Why most landing pages fail at trust Trust is the third unspoken question on any landing page. Most pages don't answer it. Three concrete moves that fix that. → (75) Vite vs Turbopack — bundler war update Vite owns the dev experience. Turbopack from Vercel is faster on large monorepos. Where each makes sense in 2026. → (76) Flutter vs React Native — practical 2026 Both shipped to thousands of production apps. The honest comparison after using both for years. → (77) Server Components two years in RSC was the future of React in 2024. By 2026 it is split between believers and pragmatists. Where it actually helped. → (78) WatchOS apps — when worth building Apple Watch is a successful platform with 200M+ users. Most apps don't need a Watch version. Some absolutely do. → (79) Mobile onboarding design 2026 The 5-screen "swipe to learn" onboarding is dead. What replaces it and why. → (80) Design systems in 2026 — what stuck Five years after every team built one, design systems split into two camps: living and abandoned. Why most fail and what survived. → (81) Mobile paywall design patterns When to show, what to show, how to compare tiers. The paywall is the single most-tested screen in any subscription app. → (82) Why we left Vercel — cost and lock-in Vercel made deployment magical. Then the bill landed. We moved 8 projects to Cloudflare Pages, saved 70%. → (83) RevenueCat vs Superwall Two tools that solve different parts of the subscription problem. Many apps use both. Here's when each. → (84) Dark UI contrast — getting accessibility right Pretty dark UIs often fail WCAG. Two ratios to remember and one tool that catches mistakes before users complain. → (85) Mobile deep links in 2026 Universal Links, App Links, Branch, Firebase Dynamic Links — what's alive and what to use in 2026. → (86) Cloudflare Workers — when to bet your stack Workers run JavaScript at the edge in 200+ datacenters. Latency drops, cold starts disappear. Hidden costs and pitfalls. → (87) App size and bundle in 2026 Bigger app = fewer installs. Especially in regions with slow internet or expensive data. What to cut. → (88) Capacitor vs React Native — when each wins Capacitor wraps a web app in native shell. React Native compiles to native. Different tools for different problems. → (89) SQLite in production — the renaissance For years SQLite was "just a dev database". Litestream, Turso and rqlite changed that. Real production stories. → (90) Android 15 foreground service quirks Each Android version tightens background work rules. API 34 added types. API 35 added permission tiers. What broke and how to fix. → (91) iPad app design 2026 — what changes Stretched iPhone app on iPad is bad UX. Real iPad app uses split views, multitasking, Apple Pencil, keyboard shortcuts. → (92) Next.js 15 Server Actions in production Server Actions reduce CRUD code by 30-50%. They also have edge cases. What we found shipping them at scale. → (93) TanStack Query as state manager TanStack Query (formerly React Query) was built for server state. Many teams use it as their primary state tool. It works. → (94) Realtime: WebSocket vs SSE vs polling Three ways to do live updates on the web. Each has its niche. Most teams pick the wrong one for the job. → (95) Astro 5 — should you migrate from Next.js Astro positioned itself as content-first. By v5 it has Server Islands and full DB integration. When it beats Next.js. → (96) tRPC in production — when it works tRPC gives type-safe APIs without OpenAPI or codegen. In a TS monorepo it's magic. In microservices it's the wrong tool. → (97) Web app i18n stack 2026 Localizing a web app is more than translating strings. Dates, numbers, currencies, plurals, RTL, SEO — all different. → (98) WebGPU finally lands — what to actually build WebGPU shipped in Chrome, Edge, Firefox and Safari by late 2025. Now what. Browser-side ML inference is the killer app. → (99) Feature flags for web apps — what to use Without feature flags, big features turn deploys into rituals with rollback drills. What to use for small and mid teams. →