My new article on Docker: "You Want Microservices, But Do You Really Need Them?"

Covers reversal stories from Amazon Prime Video, Twilio Segment & Shopify, insights from GitHub's ex-CTO & GraphQL's co-creator, and explains why function calls often beat network hops.

www.docker.com/blog/do-you-...

I'm building a small AI agent (or let's just say a smart little tool) to help founders, researchers, and investors stay on top of the latest advances in their field. 👀

Here's what it does: 👇

If you’ve been building AI agents, you’ve probably hit this point - tool integrations start getting tangled, and nothing feels clean anymore.

That’s where the MCP vs Function Calling question really starts to matter. Here is an article I wrote comparing these two.👇

www.descope.com/blog/post/mc...

Just wrapped an awesome MCP workshop (Build A Simple MCP Server) with 80% turnout on a Saturday morning. ☕️

The questions, the energy, the "aha!" moments when they got it working with Claude Desktop—you need to experience it.😇

Absolutely loved doing this live session!🙌

I’m planning a free online half-day workshop on MCP (Model Context Protocol). I’ll go over the basics and walk through building a simple MCP server.

If you’re interested, drop a 👋 and I’ll let you know when it’s scheduled (likely early October).

Here is my article evaluating 7 leading agent builders across code-first & low-code options.

A practical guide for developers and tech leaders exploring how to build AI agents of varying complexity and scale. 👇

www.descope.com/blog/post/be...

Building AI Agents? 👨‍💻

AI agents are hot, but how do you choose the right platform to build them?

From code-first frameworks like LangChain to no-code/low-code platforms like Botpress, the options are multiplying and each has strengths you need to know.

What’s Agentic AI? 🤔

It’s a shift from one-off prompts to autonomous systems: AI-powered agents that decide, act, and adapt toward goals with minimal guidance.

Curious how Agentic AI actually works (beyond the buzzwords)? Read on....

Here is my beginner-friendly article for Redpanda about understanding Apache Kafka architecture, messaging patterns and typical use case.

If you're building event-driven systems, this might save you some pain 👇

www.redpanda.com/guides/kafka...

Apache Kafka is not just a message pipe.

It’s more like a DVR for system events than a walkie-talkie. A distributed, durable log built for replay, ordering, and scale.

Used right, it’s brilliant. Used wrong, it’s... a very expensive queue.🤷‍♂️

Pune’s H. V. Desai Eye Hospital, with AI experts from MIT-WPU & Symbiosis, has developed a computer vision tool that detects retinopathy in premature babies with 95% accuracy—helping prevent the vision loss commonly seen in such cases. #AIforGood

It is fascinating to see what all this seemingly simple loop can achieve.

AI agents are going to be fun.🤩

“Everybody gets so much information all day long that they lose their common sense.”
— Gertrude Stein

It's important to distance yourself from information overload at times. 😇

Here is another example of using an AI tool (Lovable) to build something non-trivial, from a Reddit post.

The OP also draws attention to AI lies (hallucinations) and security issues.👇

I wrote this in an article about AI coding recently.

> If you don’t fully understand the code AI is producing, you might end up in serious trouble faster than you expect.

> AI doesn’t inherently think about security risks. Unless explicitly prompted,it won’t remind you...

A real-life example.👇

Have you ever set stock price alerts? Instead of constantly checking, you get a notification when your stock hits a certain price.

That's exactly how webhooks work–instead of repeatedly looking for updates, the receiving system is automatically notified by the sender. 🔔