Netlify

There are a lot of strong programmers who are deeply into music, and I don't think that's random. When I think about classical music and jazz, what strikes me is how we have to hold a very long abstract structure over time in our head and find our way through it. The classical composers were kind of like the first programmers. They had to put down an algorithm on paper - a score - and imagine how it would work when a group of people performed it. They had to imagine the whole thing over time, in their head, before anyone played a note. It's very similar to holding the abstract structure of a large computer system in your head. Being able to focus on it, reason about it, see how the parts connect. I think it's a very similar part of the brain. But there's a different piece too. As part of studying musicology, I had to conduct. Standing in front of a hundred-person choir, directing them through a piece of music. Way out of my comfort zone at first. Just the exposure of standing in front of that many people and having to lead them. Later I had the fortune of seeing some of the best conductors in the world up close, giving master classes to orchestras. And it's fascinating. It's an extreme ability to focus a whole group of people and get them aligned with a vision. Not just through what you say, but through gestures and nuances. Both the storytelling and the abstract storytelling of taking them through something. This translates directly to leadership. Getting both our own company and the world around us to believe in something - to see a direction before it's obvious - requires that same ability to conduct without controlling every note. We don't get to play every instrument. We set the tempo and the dynamics and the phrasing, and we trust the people around us to bring the music to life. I think that's probably the most transferable skill from music to building a company - not the technical precision, but the ability to make a group of people sound like they share one mind.

Claude Desktop is becoming a starting point for a lot of developers. This demo from Sean C Davis shows how you can go from prompt → website → live deploy with Netlify without leaving the chat interface. Fast workflows matter. Especially when you’re experimenting, prototyping, or shipping small ideas quickly. https://lnkd.in/gN-WGTCE

One of the most interesting points from Mathias Biilmann Christensen in this conversation: Developers often say they prioritize scalability, security, and performance first. But in practice, the onboarding experience and speed to shipping are usually what determine whether they adopt a tool at all. Netlify’s early insight was that reducing friction mattered more than leading with infrastructure specs. “How fast can we get people from not having something on a URL to having something on a URL?” That was the real product differentiator. cc Evil Martians

What does it actually take to ship a real-time AI chatbot today? The traditional answer: model inference setup, streaming logic, authentication, backend functions, deployment configuration. Hours, sometimes days, before you see anything working. In Nahrin Jalal's new tutorial, the answer is one prompt. Agent Runners generates the full stack from a single instruction: ✨ Netlify AI Gateway for inference ✨ Edge Functions for real-time streaming ✨ Serverless Functions for backend logic ✨ Managed auth with account creation and email verification ✨ Instant deployment on the Netlify platform The post walks through the exact prompt that produced the most reliable results, the CLI equivalent for terminal-first workflows, and follow-up prompts to add chat history, guardrails, a model selector, and a custom persona. The shift Nahrin captures: you're no longer assembling infrastructure, you're iterating on outcomes. Read the full tutorial: https://lnkd.in/dQ8fSxvp

“It’s not what people tell you. It’s what they do.” Mathias Biilmann Christensen on why product teams should pay closer attention to behavior than compliments: Most people won’t tell you your product is bad in a live demo. They’ll be polite. They’ll say it looks interesting. They’ll thank you for showing it. But the real validation comes afterward: * Do they return to the product on their own? * Do they actually start using it? * Do they adopt it without needing constant guidance? Because future users won’t get a founder-led walkthrough. That’s the signal that matters most. Shoutout to Evil Martians for the conversation!

The Nuxt team has disclosed four security vulnerabilities, including a route middleware bypass that affects all Nuxt apps using .server.vue pages with middleware-only authentication, regardless of hosting provider. Two of the four CVEs have limited or no impact on standard Netlify deployments. The other two are framework-level issues that require upgrading. We strongly recommend upgrading to nuxt 3.21.6 or 4.4.6 as soon as possible. Our changelog has the full breakdown of each CVE and what it means for Netlify users. https://lnkd.in/ggEYUeyg

“What you’re seeing right here exists in production.” Domitrius Clark walks through how Netlify isolates every agent run into its own environment and database branch, allowing teams to safely experiment without affecting live user data. That means you can: • test multiple features simultaneously • iterate in parallel • safely explore changes with AI agents • avoid accidental production issues Instead of one shared staging environment becoming a bottleneck, every experiment gets its own isolated workspace. This is a huge shift in how teams can build, test, and collaborate with AI-assisted workflows.

Google's Gemini 3.5 Flash is now available in Netlify AI Gateway with zero configuration required. Use the Google GenAI SDK directly in your Netlify Functions. No API keys to manage, no auth to wire up. The AI Gateway handles caching, rate limiting, and authentication automatically. Available for all Function types. Read the changelog: https://lnkd.in/gqvzanys

“Scale to zero” sounds technical, but the idea is simple: Don’t pay for infrastructure you aren’t using. Karthik Puvvada and Elad Rosenheim explain how Netlify’s new database experience automatically puts inactive databases to sleep to reduce compute costs, then spins them back up on demand in under a second. The goal is giving developers flexibility between: • lower cost experimentation • predictable production performance Once an app is live, teams can choose to keep databases always-on for faster response times and stable production workloads.

Our CTO Dana Lawson is speaking at Tessl's AI Native DevCon on June 1. Her talk: Built for Humans. Now Agents Are Here. Netlify's platform was designed around a human loop. A developer pushes code, scans the build logs, checks the deploy preview, decides it's good. Every surface assumed someone was paying attention. Then agents started driving it. Logs written for human eyes became noise to a parser. Deploy previews built for visual review couldn't tell an agent whether the result was correct. CLI output optimized for readability turned into ambiguity. The branch workflow that feels intuitive to a developer became a decision tree with no clear answer. Dana is going to walk through what concretely broke, what Agent Experience (AX) means in practice, and why the scarce resource in an agent-assisted world isn't code. It's the judgment to build systems agents can actually work with. If you're thinking about how your platform holds up when agents are the ones using it, this one is worth the time. Register: https://tessl.io/devcon/