🦸 Community Superhero

Our community superhero this week is Peter Lorent, senior AI/ML specialist SA at Philips. Peter is an active member of the Believe in Serverless community where he regularly gives insights into building software with and without AI capabilities. He has a lot of experience in the industry and is always happy to help when he can. I learn a lot from him on a weekly basis. Thank you for everything you do, Peter!

💯 Spotlight

I caught on to a community collaboration last week that made me go “that’s what I’m talking about!” as soon as I read it. Sylvain Bruas shared an MCP server he wrote to help plan which AWS re:Invent sessions to attend. That’s cool by itself, but what really got me excited was that he uses Raphael Manke’s unofficial re:Invent planner as the API that powers it. Community on top of community! This right here made my week. Thank you both so much for contributing all this hard work to make it easier for the rest of us

🔥 My Favorite Content

AI agent tools are APIs. That was the message Andres Moreno and I gave over and over and over again in our livestream last week that discussed enterprise-grade agent tool design. This stream went heavy on API design principles, emphasizing patterns we already know, but have renewed importance with agents quickly becoming the primary consumers of public APIs. We talked a bit about MCP, the dispatcher pattern for governance checks, and things to consider when defining your API specs for tool consumption. If you happen to be in or around Newcastle upon Tyne this week, I’m speaking in depth on this topic at the AWS North Community Conference on Thursday. Come by, say hi, and learn a thing or two about API design in the AI era.

Nothing makes me struggle more than trying to build observability dashboards. There’s so much I don’t know about meaningful metrics and how to evaluate them correctly. I was particularly interested in Daniela Miao’s post last week where she explained 5 metrics that predict cache outages and why they aren’t the ones you’d think. It’s a neat article with great explanations and fascinating mental reframing for monitoring what matters.

Eric Johnson recently chatted with the team from AntStack, and I loved one of the little clips they pulled out from their conversation. Eric talks about how generative AI is just another serverless workload. It makes so much sense and completely validates (to me) so much AI bleed into this newsletter. Great take, Eric!

💡 Tip of the Week

In a smart move from AWS, next year they will be moving away from the XRay SDK and daemon and moving toward OTEL. I love seeing them make an effort on normalizing observability 🤩

🐣 New Releases

Open AI released their Apps SDK, letting builders create agentic workflows by adding logic and UI into their apps. Seems like a cool development here.

AWS introduced Amazon Quick Suite, and the best I can tell it’s a research agent and general assistant designed for the enterprise. Seems interesting!

Last Words

I’m in England this week! If you’re in Manchester or Newcastle and want to meet up, send me a message and we’ll get together. I’ve been here a few days now, so I’m just starting to get over the jet lag and ready to see some friends.

That’s my take on the week, but what’s yours?

What did I miss? What made you nod along (or 🙄)? Hit reply if you’re reading the email. Prefer socials? Ping me on Twitter, LinkedIn, or email.

Happy coding!

Allen