🦸 Community Superhero

Our community superhero this week is Sylwia Laskowsa, senior engineer at Atos, and a wonderful blogger and speaker. She regularly shares hands-on experiments with WebGPU, WebAssembly, AI agents, and (my favorite) bizarre JavaScript behavior. She’s also an active contributor to the DEV community, spending time helping other developers and sharing what she’s learned. If you enjoy technical content with a little personality and working examples, give Sylwia a follow. Thank you for everything you do, Sylwia!

💯 Spotlight

Over the years, we’ve seen lots of content evangelizing serverless architectures and the “right” way to build them (especially in this newsletter). I really appreciated a post from Jesús Hernández that acknowledged the golden era of serverless (around 2022), but also goes into the growing pains of it with the prominence of AI. Jesús talks about recent evolutions they’ve made and productivity boosts to keep their builds fast and confidence high. He’s setting us up for a series on the topic, and I’m very much looking forward to it.

🔥 My Favorite Content

I haven’t tried Lambda Web Functions yet, but it’s been on my short list of things to try. After reading Danielle Heberling’s post about how it’s the Lambda she’s wanted since 2017, it’s moving up even higher. I like how her article describes the differences between Web Functions and “normal” Lambda (now known as Event Functions), and she explains a bit about what she built as a first look into it. This isn’t a heavy blog post about engineering principles, but rather a delighted reaction to something that’s genuinely useful. Love it.

Taking Web Functions a step further, Massimiliano Angelino published an article where he used them to host an MCP server behind AgentCore Gateway. For whatever reason, this wasn’t a use case that had originally occurred to me, but it makes sense and Massimiliano does a great job explaining how to do it and how to secure it with Cognito. The article takes you through a simple example that you can easily extend for your projects.

You know those super cool agent skills you build and are excited about, then your coding agent forgets to use them? Yeah, same. And apparently same for Kenta Goto. Kenta published an article last week introducing an open-source tool he built that makes sure your agents are using the skills you want them to. It’s an interesting idea that I really like and think it has a great place in enterprise or top-down AI mandates. Unfortunately the content feels heavily written by AI, but the tool seems to have merit. Worth the read and a POC.

With all the excitement of AI advancements, many of us have been wearing ourselves unintentionally thin. It’s something that is becoming easier and easier to do, and as Brian Tarbox puts it, it’s like butter scraped over too much bread. In his article last week, Brian makes the argument that we’re not leaving ourselves with enough slack, and trying to operate at 100% all the time. We don’t let our servers run at 100% capacity, so why would we do that to ourselves? This is an honest write-up that’s sobering and reminds us that we all have limits. Bonus points to Brian for the Lord of the Rings references throughout the content.

💡 Tip of the Week

I’ve spent the last several weeks vigorously building a “Dungeons & Dragons” style adventure game for kids, and the amount of code, graphics, and audio I’ve been pushing is absolutely ridiculous. While the code velocity has been unprecedented, bottlenecks have been showing up elsewhere, like in review time when I make big changes or in QA time (waiting for my daughter to get home from school to try it out 😜). I read an AI Engineering Report from Faros that validated I’m not the only one experiencing this.

Last Words

Woah, we’re only 4 weeks away from re:Invent! That snuck up quickly. Who am I going to see there??

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