Why AWS Went Down & What’s New in Next.js 16

Plus: OpenAI founder says AGI’s still a decade away & Perplexity releases new work guide

Hello from San Francisco! ☀️

I’m checking in from Next.js Conf with our CTO, Jakub Dakowicz. Between talks from industry legends and way too much coffee, I’ve been trying to keep up with everything happening this week.

Chris Lojniewski and Jakub Dakowicz from Pagepro at Next.js Conf in San Francisco, holding Sanity-branded coffee cups and an AWS immunity shot, smiling outdoors in front of a mural.

And what a week it’s been! Next.js 16 is finally out. Meanwhile, AWS went down for 15 hours, reminding everyone why observability isn’t optional. Meta’s new AI chief is shaking things up after 600 jobs were cut from its AI division, and Amazon’s preparing to replace half a million jobs with robots

In case you missed it, OpenAI launched its own agentic browser, but security experts suggest you might want to hold off on downloading it just yet. Speaking of OpenAI, one of its founders, Andrej Karpathy, shared why he thinks true AGI is still a decade away.

Sounds good? Then grab your own coffee and enjoy the new issue of Frictionless!

In the Queue

Reduce Friction

A grid of screenshots showing service status spikes for major platforms — AWS, Amazon, Reddit, Snapchat, Venmo, Roblox, and others — visualizing the impact of the recent AWS outage.

AWS Outage and Why It Again Proves Full-Stack Observability is Non-Negotiable

A 15-hour AWS outage in the us-east-1 region took down more than 140 services after one DNS glitch spiraled through DynamoDB. Even the biggest platforms weren’t safe, but teams with proactive monitoring caught the issue before AWS itself did. Read why visibility across your stack can make all the difference.

The Price of Mandatory Code Reviews

After studying data from 400+ companies, Anton Zaides found what many suspect: mandatory reviews slow teams down. Skipping them entirely makes the code nearly twice as buggy, though. The million-dollar question is, how do you keep speed and quality?

Engineering Leader’s Guide: How to Become a Great Coach and Mentor

Managing projects shouldn’t be an engineering leader’s only focus. The best ones invest time in growing their teams by asking the right questions instead of giving all the answers to save time. It helps developers think for themselves and stay motivated, no matter how many deadlines they’re chasing.

5 Tips to Succeed with Stakeholder Management

Getting everyone on the same page can feel impossible, but it doesn’t have to be. This guide shares five ways to manage stakeholders without endless calls or watered-down decisions. It’s a quick read that can save you a lot of friction on your next project.

Deepen Your Expertise

The official black-and-white Next.js 16 logo featuring the text “NEXT.JS” and the number “16” in a rounded rectangle — the branding for the latest version release.

Next.js 16: Cache Components, Stable Turbopack, and AI DevTools

Just as I arrived at the Next.js Conf, Vercel released the latest version of Next.js. Turbopack becomes the default bundler for faster builds, caching gets smarter, and DevTools now come with AI support. A solid step forward for anyone building with React and Next.js.

Inside the Breach that Broke the Internet: The Untold Story of Log4Shell

In 2021, a single vulnerability in Log4j, a tiny Java library, nearly brought the internet to its knees. Dubbed Log4Shell, it exposed everything from Fortune 500 servers to Minecraft worlds. See the human side of that crisis through the eyes of the one maintainer who suddenly found himself responsible for fixing half the internet.

NextJS Performance Optimization in 10 Steps

I’ve spent a lot of time helping teams fix slow, expensive Next.js apps, and this guide sums up what works best. It covers 10 practical ways to boost speed and stability without rewriting your whole stack. If you care about smoother user experiences and lower Vercel bills, it’s a great place to start.

The Country That Broke Kotlin

Who knew a single letter could cause so much chaos? A tiny quirk in the Turkish alphabet ended up breaking Kotlin’s compiler and confusing developers for years. See how one small detail in language rules spiraled into a global headache for engineers, and what it teaches us about the messy intersection of human language and code.

AI Corner

A still from an interview with Andrej Karpathy, where he gestures mid-sentence. The text on screen reads “LLMs are a different kind of intelligence.”

Andrej Karpathy - AGI is Still a Decade Away

Andrej Karpathy, one of OpenAI’s founding fathers, chats about the state of AI and where it’s heading. He argues that AGI isn’t around the corner, that reinforcement learning is “terrible but everything else is worse,” and that AI progress will blend into the same steady 2% GDP growth we’ve seen for centuries.

Perplexity at Work: Official Guide to Getting More Done

Perplexity released a guide showing how to use AI to simplify your work instead of adding more of it. It’s built around three stages: first, blocking distractions, scaling your abilities, and finally, getting tangible results for your team or business. It comes with practical examples too!

Dane Stuckey (OpenAI CISO) on Prompt Injection Risks for ChatGPT Atlas

OpenAI’s new Atlas browser aims to make ChatGPT your main way of browsing the web. But as Simon Willison points out, giving AI that much control comes with new risks. He digs into how OpenAI’s CISO plans to defend Atlas against prompt injection attacks and keep users like us safe.

Meta Lays Off 600 from ‘Bloated’ AI Unit as Wang Cements Leadership

After Llama 4’s underwhelming debut, Meta’s AI empire is getting a full reset. 600 roles are gone, and insiders say it’s more about power than payroll. Alexandr Wang, brought in after Meta’s $14B Scale AI deal, is tightening his grip as cuts hit legacy FAIR and infrastructure teams hardest.

Just Cool

A worker in an Amazon warehouse operates near bright yellow bins and automated machinery, illustrating Amazon’s growing use of robotics in its fulfillment centers.

Amazon Plans to Replace More Than Half a Million Jobs With Robots

Leaked internal documents show Amazon’s bold new plan to automate 75% of its operations, potentially replacing more than 600,000 human roles over the next decade. Amazon says it’s reimagining jobs, not cutting them, but the scale of automation is bound to affect local communities.

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.

What do you think of today's email?

Your feedback helps me improve Frictionless

Login or Subscribe to participate in polls.