Writing

The exciting promise of AI is that everyone can code. But packaging, sharing, deploying, and operating is a steep step function. Astro AI is working to systematically solve this for a much broader definition of builder.

Introducing Astro AI: 10x faster agent experimentation, centralized management with guardrails, blueprint sharing, and data source visibility. Live Minecraft AI battle royale demo at KubeCon26 EU.

Astro AI is a production-grade control plane for deploying AI agents as operational assets rather than experiments, with standardized packaging, governance, observability, and lifecycle management.

Two agents. Same model. Same prompt. One connected to Postman's API Network. The other? On its own. The difference: one built the entire app without errors. The other failed three times. Context is everything.

It was surely always true that no matter your business, it wasn't safe from someone ambitious, determined, and hungry enough to come eat your lunch. It's certainly true now with AI and APIs. Stay hungry.

Most people still think of Postman as just an API client. In reality it's a platform for 40M+ developers: discovery, testing, automation, and AI-powered workflow building via Flows and Agent Mode.

At KubeCon, I took a prompt from Spotify on the spot, entered it in Postman agent mode, and it built a flow that finds the top 20 playlists in Atlanta. The energy around Flows and API-first AI was undeniable.

I use Postman agent mode constantly to understand unfamiliar APIs and test suites. The best part: it brings the full power of gen AI to where I already spend my time. No context switching to ChatGPT.

Programming is undergoing a fundamental shift, from writing explicit logic to orchestrating AI-driven solutions through APIs and visual interfaces. The developer role isn't disappearing, it's transforming.

How to build a production AI app (natural language to terminal commands) using Postman Flows, mock servers, GitHub, Stripe, and gen AI models. No traditional backend required.

A RISC-y analogy for the emerging cloud computer and what serverless computing means for the future of software development.

A deep look at security considerations for serverless functions and what developers need to think about when building on serverless platforms.

What serverless computing promises to developers, and what developers must understand to use it effectively.

How to compose serverless functions into higher-order applications using the Composer programming model.

Using the IBM Cloud Shell to profile and debug performance issues in serverless functions running on OpenWhisk.

OpenWhisk supports a wide range of programming languages. Here is how to write and deploy serverless functions in the language you already know.

Serverless functions and APIs are two sides of the same coin. Understanding their relationship unlocks a more powerful model for building cloud applications.

How to move beyond individual serverless functions and compose them into full applications with control flow, error handling, and state.

Building HTTP endpoints and web APIs directly from OpenWhisk serverless functions, without managing servers or routing infrastructure.

How to precompile Swift actions for OpenWhisk to get dramatically faster cold start times and better performance.

Extending OpenWhisk with Docker-based actions to run any language or runtime, including Golang, in a serverless environment.

Advanced techniques for debugging serverless functions locally before deploying them to OpenWhisk in the cloud.