The Weekly Dev's Brew Is Pivoting

31 newsletter issues. 16 podcast episodes. A year of curating JavaScript ecosystem links. And somewhere around issue 25, I realized I was bored writing it.

Not burned out. Bored. That's worse.

What stopped working

The Weekly Dev's Brew started as a JavaScript newsletter. Weekly links, framework releases, ecosystem takes. Solid idea. Except there are already a dozen good newsletters doing exactly that, and I was fighting over the same links every week.

Open rates were fine. Nobody complained. But the issues people actually replied to? Those were the ones where I'd tried something myself and wrote about what happened. Not "Angular released v19" but "I spent a week using AI to review my PRs and here's what it caught."

The link roundup format has a ceiling. I hit it.

Why AI workflows

I've spent the last year watching developers around me change how they work. Not in the "AI will replace us all" way. In the boring, practical way: an agent writes the first draft of a test suite, someone uses Claude Code to scaffold a migration, a teammate sets up Cursor rules for their monorepo.

The coverage of this is mostly garbage. Either it's a company blog pretending their tool solves everything, or it's a Twitter thread from someone who tried Copilot once in 2023 and decided AI is useless.

There's almost nobody doing what I want to read: here's what I actually used, here's what actually happened, here's what it actually cost. The "I tried it so you don't have to" angle.

That's the newsletter I want to write.

What changes

The newsletter gets a new structure. Five sections per issue:

• The Experiment takes up most of the issue. One real experiment per week. I compare tools, I track costs, I show prompts, I report failures. No hedging.
• The Workflow is a concrete, copy-paste-able breakdown of one AI-assisted task. Not tips. Steps.
• The Take is an editorial opinion where I pick a side. Short.
• The Links stay, but every link gets my actual opinion attached, not just a title.
• The Community Spotlight rotates between reader workflows, questions, and challenges.

The podcast goes biweekly. Same vibe, different guests. Practitioners building AI dev tools or using them seriously. Seven segments, structured to produce short-form clips. No product launch press tours.

The JetBrains thing

I should address this. I work at JetBrains. JetBrains makes AI developer tools. That's a conflict of interest and I'm not going to pretend otherwise.

The deal is simple: when JetBrains tools are great, I'll say so. When they're not, I'll say that too. When a competitor does something better, that goes in the newsletter. If readers can't trust the content, none of this works.

What I'm losing

The JS community niche was valuable. Specificity builds audience and I'm giving that up.

The link roundup was also fast to produce. The new format is a lot more work per issue. Running experiments, writing them up, tracking metrics. That's not a Tuesday evening task anymore.

Worth it? I think so. Ask me again in six months.

Starting now

The next issue follows the new format. The podcast is already transitioning.

If you subscribed for JavaScript links and this isn't your thing, no hard feelings. If you're a developer using AI tools and you want someone to actually try stuff and tell you what happened, stick around.