At React Universe 2025, I had the pleasure of sitting down with the folks from Amazon to talk about how Klarna ships multiple apps from a single codebase with confidence. I’m sure I forgot a bunch of things 😅 but hopefully this snippet paints a picture of what’s possible and what to consider when designing for this kind of setup. I believe in sharing what I’ve learned so others can build on it and make it better. If you approach this differently or have ideas to add I’d love to hear them!

Still writing tests manually? Meticulous AI is here Most teams are forced to make the tradeoff between better coverage and more maintenance effort. But top engineering teams like Notion, Dropbox, Wealthsimple and LaunchDarkly have discovered a new testing paradigm. Built by ex-Palantir engineers, Meticulous creates and maintains E2E UI tests that covers every edge case of your web app - without any developer effort, making it the only tool to improve both product quality and dev velocity. https://lnkd.in/d393z9TJ

While your team is busy getting "assisted" by AI, elite engineering orgs are completely automating their software development. Companies like Ramp and Stripe aren't just trying to type faster; they are deploying background agents that read tickets, spin up isolated sandboxes, and open pull requests automatically. The industry shift is moving rapidly from conversational coding bots to an autonomous, parallel development workforce. Is your team actually automating the SDLC, or are you just chatting with your code?

In case you wonder how other companies are integrating AI into their internal developers ecosystem. Stripe went all-in with AI generated code. They have a pretty clever setup to work in their historically grown codebase, they don't rely on agents alone, they pre-feed context and have deterministic stages to check work done. I like their term of minions here. Have a read, I encourage you to do so. https://lnkd.in/dG3CM9Rt https://lnkd.in/dmDEGR_y

Stripe Deploys Autonomous Minions to Generate 1300 Weekly Pull Requests 📌 Stripe just unleashed Minions - autonomous coding agents that auto-generate 1,300+ pull requests weekly with zero human-written code. These AI-powered bots handle everything from dependency upgrades to refactoring, all within isolated devboxes and strict blueprints, ensuring quality before human review. A massive leap from assisted coding to full-scale, self-running software engineering - even in billion-dollar payment systems. 🔗 Read more: https://lnkd.in/dirCYxSB #Stripeminions #Pullrequests #Autonomousagents #Codegeneration #Slackintegration

Software Development: Is AI Lowering the Barrier to Entry? Want to see more where this clip came from? Check out the video here: https://lnkd.in/gCkEaVNZ Subscribe to Code Commute on YouTube and Spotify for more!

👇absolutely agree the barrier of entry is lower than ever to build edge apps in fin/hr/rev/ procure ops, and most are eager to learn but don’t know how to get started. Having domain/process knowledge and knowing how to harness AI for is a key differentiator. Watch, learn, and jump in!

This hit home. What Angela Liu and Devon did here is bigger than the app itself — it’s proof that finance operators can now build, not just operate. I’ve been thinking a lot about this shift: → The bottleneck is no longer engineering → It’s problem clarity + domain judgment We’re moving from: Finance = system users → Finance = system builders And the edge is no longer technical engineering skill — it’s understanding the problem deeply enough to design the solution. This is exactly the direction I’m leaning into. Feels like we’re just at the beginning of something big. Hope to connect with more peers who are also excited on this journey! Thanks for sharing!

Your AI just wrote 1,000 lines of code in seconds...are you sure the app works? Faster code ≠ working software. If development is moving at AI speed, quality has to keep up. Dan Faulkner in Forbes Technology Council: https://lnkd.in/gCykXxcE

Your AI just wrote 1,000 lines of code in seconds...are you sure the app works? Faster code ≠ working software. If development is moving at AI speed, quality has to keep up. Dan Faulkner in Forbes Technology Council: https://lnkd.in/g_ZbvmYv