Amazon’s 30x Growth & Netflix’s Java Stack

Hello!

Every engineering leader has a moment where something that used to work suddenly doesn’t. Maybe your team’s getting slower as it grows. Maybe your MVPs aren’t giving you answers. Maybe you’ve gone “AI-first” and still feel behind.

Last issue, the most-read piece was about how Stripe manages to ship 1,145 pull requests a day, the most. 

This week, we’re continuing that theme of scalability with a piece on the seven hidden blockers that stop teams from shipping, and a peek into Amazon’s strategy for keeping productivity on track while scaling to 50,000 engineers. 

You'll also get a look under the hood of Cursor and how Netflix is using Java to stay performant.

Grab your coffee, relax, and enjoy reading Frictionless!

In the Queue

• Pen and Paper - All Devs Need?

• Future-Proofing AI

• Building Rich-Context AI Apps with Anthropic

• Guide to Sanity.io

Reduce Friction

When a Team Is Too Big

Ever notice how delivery speed tends to go down as team size goes up? Find out why the strategy of  “just add more people” often backfires on engineering teams, and what high-performing teams do differently.

This Is Why You're Not Shipping

Developers don’t miss deadlines because they’re bad at their jobs. Sometimes it’s your company that’s getting in the way. Here are seven blockers your team is probably facing and their fixes.

Inflection Points in Engineering Productivity as Amazon Grew 30x

Scaling a team from 3,000 to 50,000 engineers doesn’t happen without hitting a few snags. Amazon turned those moments into data, and that data into decisions, improving how they operate.

Deepen Your Expertise

Your MVP Didn’t Fail; You Didn’t Set It Up to Teach You Anything

We’ve shipped enough MVPs at Pagepro to know that launching one should always teach you something useful. If your latest MVP didn’t do that, this framework might help you ask the right questions and get the answers you need.

Is Sanity.io the Best CMS for 2025?

With its latest updates, Sanity has grown from a headless CMS into a Content Operating System. If you’re looking for a tool that ensures a scalable and flexible content setup, now’s the time to take a serious look at what is Sanity.

How Netflix Runs on Java

Netflix’s backend is a masterclass in using Java for large-scale systems. Learn about the Java 21+ features and architecture choices that help them stay fast and in control of a massive microservices setup.

AI Corner

Why Companies Are Really Going AI-First

Duolingo, Shopify, Klarna... all these companies are proudly declaring themselves as AI-first, but why? To answer that, we need to look at where they’re applying AI, and what those choices say about the future of engineering teams.

Future-Proofing AI: Repeating Mistakes or Learning From the Past?

Everyone’s racing to scale their AI stack, but are we making the same mistakes we did during the cloud boom? If you want to avoid déjà vu, check what to watch out for, and how to learn from the past.

9 Lessons From Cursor's System Prompt

Most AI assistants feel clunky, but not Cursor. Its system prompt is carefully engineered to guide behavior, tone, and structure. See how it works and what you can apply to your own AI tools.

MCP: Build Rich-Context AI Apps with Anthropic

Got an hour and a half to spare? Anthropic’s latest course is a hands-on intro to building context-rich AI apps using Claude and the new Model Context Protocol. Easy to follow, and full of ideas you can plug right into your workflow.

Just Cool

As A Developer, My Most Important Tools Are a Pen And a Notebook

Every day brings a new tool that promises to change how we code, but maybe we need something simpler. Remember pen and paper? They might still be your best bet, even in the age of AI.

Let’s Stay in Touch! 📨 

Do you have any comments about this newsletter issue or questions you want to ask? Drop me a message or book a meeting.