Issue 114 · 25 Feb 2026

The frontend moves fast.
This is the signal.

Every Tuesday at 7 AM — one opinionated briefing for engineers who need synthesis, not another feed.

Scroll

001 — Why

The problem isn't access to information.
It's the cost of processing it.

Hacker News refreshes every four minutes. The JavaScript ecosystem ships a new framework roughly every eleven days. Twelve browser vendors publish release notes no one has time to parse. Meanwhile, your team is waiting on a decision about whether to adopt React Server Components, whether Vite has finally made Webpack obsolete, whether the Svelte 5 migration is worth the sprint cost.

The noise is not the problem. The problem is that noise and signal look identical until you've already spent four hours on a thread that ends in a meme.

Dispatch exists for one reason: to do the reading so you don't have to, and to tell you what it means — not what it says.

114

Issues published

47k

Engineers subscribed

7 AM

Every Tuesday

"Most newsletters give you more to read. Dispatch gives you less to decide."
— Issue 108 — Editorial

002 — What

Every issue. Three sections.
Zero filler.

The Trend Digest

Three framework or tooling shifts ranked by signal strength — adoption velocity, GitHub star trajectory, job posting frequency. Not what's trending. What matters and why.

// Issue 114: React Compiler v1.0 ships
// Before: manual memoization everywhere
const Component = React.memo(({ data }) => {
  const processed = useMemo(() => process(data), [data]);
  return <View data={processed} />;
});

// After: compiler handles this automatically
const Component = ({ data }) => {
  const processed = process(data);
  return <View data={processed} />;
};

One Contrarian Take

The argument no one wants to make. Why TypeScript's Go rewrite might break more teams than it helps. Why Tailwind adoption is a hiring signal, not a technical one. Why your bundle size benchmark is measuring the wrong thing.

// Contrarian: TypeScript in Go
// The rewrite is 10x faster, but...
// Your tsconfig.json assumptions break.
// Type-only imports behave differently.
// Plugin ecosystem resets to zero.
// Migration window: 18-24 months.
// Is your team on the right side?

The Decision Frame

One stack decision framed as a two-column table: what you gain, what you lose, who should care. Built for the five-minute conversation with your CTO, not the two-hour architecture review.

// Decision: Vite vs Webpack in 2026
// Gain: 10-100x cold start, ESM native
// Lose: Plugin ecosystem (still catching up)
// Who cares: Teams on Node 18+, new projects
// Who waits: Legacy CJS modules, Nx monorepos
// Verdict: Default to Vite. Migrate Q3.

"The contrarian take is worth the subscription alone."
— Priya Nair, Principal Engineer, Shopify

003 — How

The methodology
is the product.

Every source has a half-life. GitHub stars are a leading indicator for six months, then they become lagging. npm downloads measure existing usage, not future adoption. Conference talks preview where the industry is going roughly two quarters before it arrives.

We weight sources by their predictive accuracy over a rolling 18-month window — recalibrating quarterly against what actually shipped, what actually got adopted, what actually changed hiring.

The contrarian take exists because consensus is always late. By the time everyone agrees something is good, the decision window has closed for early adopters and the risk profile has inverted for everyone else.

Source weighting matrix

High

GitHub star velocity

Relative momentum over 90 days, not absolute count

High

Job posting keywords

LinkedIn + Indeed, filtered to senior/staff roles

Medium

Conference CFP submissions

Leading indicator — talks precede adoption by ~6 months

Medium

npm weekly downloads

Useful for libraries, noisy for frameworks

Low

Twitter/X discourse

Signal-to-noise ratio approaching zero; used for sentiment only

Low

Stack Overflow survey

Annual lag makes it confirmatory, not predictive

"By the time everyone agrees something is good, the decision window has closed."
— Issue 97 — On consensus and timing

004 — Proof

47,000 engineers.
Same Tuesday ritual.

"I canceled four other newsletters after the first month. Dispatch is the only one that actually changes what I recommend to my team."

Marcus Webb

Staff Engineer, Figma

"The contrarian take section is worth the subscription alone. Three times this year it saved me from a decision I would have regretted."

Priya Nair

Principal Frontend Engineer, Shopify

"I forward the Decision Frame section to my CTO almost every week. It's the only newsletter I read before my second coffee."

Daniel Osei

Tech Lead, Linear

"Building solo means I can't afford wrong bets. Dispatch is my second opinion on every stack decision I make."

Soren Lindqvist

Indie developer, Malmö

47,214

Active subscribers · 68% open rate

Read Free Archive

005 — Archive

Read before you subscribe.
We insist.

Full archive (114 issues) →

#114

25 Feb 2026

React Compiler ships v1.0 — what it actually breaks

Manual memoization is now legacy code. We break down what the compiler handles, what it doesn't, and the three edge cases that will bite you in production.

ReactCompilerPerformance

Read →

#113

18 Feb 2026