TypeScript 7.0: the Go compiler and the end of slow builds
The Go native compiler landed on 8 July: full builds 8 to 12 times faster and the first error in the editor after 1.3 seconds instead of 17.5. The benchmarks, the breaking changes, and how to schedule the migration.
INP one year on: how we keep interaction under 200ms
Interaction to Next Paint replaced FID as the responsiveness metric. A year later we know what actually moves the needle and what is theater.
PostgreSQL 17 in production: the features we actually use
Since 17 hit LTS we ran it on our largest projects. What the upgrade was worth and what stayed on paper.
Next.js 16 and React 19 six months in: what we actually used
Server Actions taking center stage, partial prerendering finally stable, and a build pipeline making less noise. Verdict in May 2026.
Web performance in 2026: INP has grown, LCP has not
Early 2026 Chrome UX Report data is clear: INP is the metric separating fast sites from slow ones. What it takes to stay under 200ms.
PostgreSQL 17 in production: three features we liked
Postgres 17 shipped in September 2024. Six months after upgrading our projects, three things that really matter.
Cloudflare Workers: the real 10ms CPU limit and how to live with it
Workers are powerful but strict. 10ms CPU per request — not wall time — disciplines how you write code.
Prisma 5.10 and relationLoadStrategy: a join in the right place
Prisma 5.10 brings relationLoadStrategy, a preview flag that lets you pick between subquery and join for relations. For anyone fighting hidden N+1, it is fresh air.
Tailwind: how many megabytes of JavaScript do we really save?
Tailwind kills CSS-in-JS and a chunk of JS runtime with it. Let us quantify the saving on three real projects.
Lighthouse 11 and the new Core Web Vitals: what changes for optimisers
PWA score gone, INP replacing FID, accessibility tightened. A small revolution in our audit reports.
PostgreSQL 15 in production: the verdict after six months
MERGE, faster logical decoding, better sort performance. What we actually used and what stayed on paper.
