I'm playing a video game with my 7 year old and when we entered the Amazon she laughed. I asked why she laughed and she said "because that's a store" 🫠🫠🫠

I've always been a backend engineer, but nothing makes an API or a service feel more real than the first time you see a functioning UI on top of it 🤩

In a recent episode of the #cacheit podcast, Khawaja Shams and Kyle Davis spoke about some of the exciting new possibilities with multiple database in cluster mode in @valkey_io 9.0 (among other things) 🔥 Full episode 👉 https://t.co/OrKhAX9jgB

Anyone else see the irony in people saying there are lots of security issues with MCP yet still freely using it? Don't get me wrong, it's a fantastic protocol - not without issues - but it's a little funny to see the popularity of local MCP servers from unverified publishers.

When I first learned about multi-agent collaboration, I thought it was going to upend how we build things. Now that I have a few months experience building autonomous agents, they honeymoon period is over and I'm favoring "agentic workflows" rather over agent tool "bingo" 🤣

An unexpected side effect of using agentic code editors - I'm way more excited to write code than I have been in quite some time. I learn so much about problems that would have taken me weeks... in just a couple of hours! I love being unblocked in minutes rather than days 💙

Prompt injection is like SQL injection for AI, except it's way easier to do and pretty effective (right now). If you aren't using guardrails in your agents, add them today! https://t.co/VQvTie28XG

Your API don't HAVE to take the exit ramp into chaos. Join me and Shashank Awasthi from @getpostman on November 6 at 10am PT to learn how API governance keeps your workflows mature, functional, and fast 😃 RSVP here👉 https://t.co/Zb3zajrGPl

The Null Check show with @AllenHeltonDev and me has a new home on YouTube!! We are currently going through each of the 6 design principles for building AI agents for the enterprise. Subscribe to our channel!! https://t.co/7ImwjGXKev

There's so much more involved in AI agent design than "throwing the book at it" and hoping it figures out what to do. Here's a clip from a recent stream I did with @andmoredev talking about instruction dilution. #ai #aiagent #promptengineering https://t.co/tGTtcViuMz

When building an AI agent, you must decide whether to provide all the context upfront or use tools to enrich the context. Which is better? .. Well .. it depends! In this clip,@AllenHeltonDev and I talk about the pitfalls of using one over the other. https://t.co/IbZuIZxlf6

What is the “lost in the middle” effect and how can you avoid it? @AllenHeltonDev dropping some knowledge bombs right here 👇 https://t.co/Us725qtoFO #ai #agent #til #aws

Tried my hand at making a comic today to describe why you shouldn't rely on RAG to do specific data lookups. Use it for finding similar things or summarizing. Use tools for exact data. #ai

💫 @momentohq boosted API engagement by 46% after publishing on the Postman API Network. ✅ More API discoverability ✅ Faster developer onboarding ✅ Less time spent on support APIs developers can find and trust get adopted faster. Case study 👉 https://t.co/noiOhATU4t

End-to-end with A2A - let's see it work!

I hope to see you there! This has become a passion project for me. I've learned so much about agents, MCP server design, A2A, and everything in between and I'm excited to share 🤓

Example code we're covering: https://t.co/ROSJsAhfFI This repo uses @Cloudflare to host MCP servers, limited-scope AI agents that consume those servers remotely, and hosts an agent orchestrator that can call on any of our agents based on the prompt.

NPM package that's ready for use: https://t.co/fSEF8Gu8UN This offers an abstraction on top of A2A that uses Momento to store task data, stream events, and register agents. It's designed to support serverless agents wherever you run functions - @awscloud, @Cloudflare, etc..

Event on LinkedIn: https://t.co/jOT9M9nwvz But I'll also be streaming live on X! I just haven't figured out a way to let people RSVP for that 😅😅

I see a lot of people talking about A2A. But I don't see many people actually using it. In fact, I'm not sure anybody is. Why do I think that? The friction to build with it right now is... 😬 But I've built agents with it. In fact, I've built a pretty solid abstraction on top