Our serverless superhero this week is Darya Petrashka, senior data scientist at SLB and AWS Community Builder. Darya spends an inspiring amount of time educating others on AWS and AI - and she’s great at it! Be sure to follow her to keep up with her wonderful sessions and incredibly valuable insights. Thank you for everything you do, Darya!
This week’s spotlight goes to an article that hits really close to home. Annie Vella wrote a masterful piece called The Software Engineering Identity Crisis that shows us exactly what’s happening to our industry with generative AI. She makes a case for software engineers gradually turning into engineering managers with the surge in code-literate LLMs, which while extreme, actually has a lot of merit to it. This one will make you think a bit.
In case you missed it like I did, the Lambda Powertools team recently launched log buffering. What is log buffering, you ask? Let me point you straight to Andres Moreno’s post from last week on exactly that. Andres shows you exactly how to use log buffering and what happens to your logs under various real-world scenarios. This is something we’ve needed for a long time, and I’m super happy to see it hit production. Great work on this explainer, Andres - it’s super helpful!
If you’ve been on the internet in the past couple of weeks, you’ve undoubtedly seen the term “vibe coding”. Don’t worry, you don’t have to be too afraid to ask what that means 😂 Carlota Soto did a great job last week explaining the term but also introducing a term of her own - tab coding. Ultimately she’s talking about using AI code editors and warning us about how they are changing our habits. She’s not wrong, and makes some good distinctions in her article.
Spring into Event-Driven Architecture! The Serverless Guru Spring Hackathon is now open for registrations—free, fully online, and with a $9000 prize pool. Buzzing around from April 25th - May 18th; join today and let your ideas bloom! Sponsored
The king of event-driven architectures, David Boyne, had a talk posted online last week on how to navigate complexity in EDA. Everything David does around EDA is gold, and this hour-long session is jam packed with knowledge everyone from beginners to years-long practitioners should know.
I’ve shared posts from around the community on MCP (model context protocol) the last several issues. It’s quickly emerging to be the widely adopted standard for AI agents to communicate with 3rd party systems. But there’s a lot of misinformation and false assumptions on it. I love Kevin Swiber’s take on it. They shared two articles on the topic last week about how MCP is the ultimate API consumer and what it means now that OpenAI embraces it. I have a feeling we’re going to see an intense focus on this protocol in the upcoming months.
Lastly, we had possibly my favorite livestream in a while. Andres and I walked through creating a multi-agent collaboration supervisor in Amazon Bedrock. We started with 4 working agents representing different characters in the tv show The Office, and walked through the AWS console to create Michael Scott, the loving boss who delegates tasks across his employees to plan the perfect party. We focus a lot about practical use of agents, how they work, and clear up some misconceptions about them. It was a fun session highlighting this incredible feature. Bonus - all the source code for that project can be found here.
I don’t normally choose myself for the tip of the week, but I feel like something I said early last week is important enough to share. One thing I love about our tech community is the thirst for knowledge we all have. Never satisfied with what you know now, many of us are constantly learning new things. But starting to share what you learn can be intimidating - I know it was for me! Just remember - we’re all here to help, not judge 💙
Amazon DynamoDB Streams now supports AWS PrivateLink, so you can invoke a streams API from within a VPC!
Pulumi announced the next-gen of Pulumi Components, which allows consumers to create components in one programming language but use them in whatever language you want. They’ve also simplified how components are shared, made them easier to consume, and added YAML support.
Cloudflare announced that you can build and deploy remote MCP servers on their infra. That’s a big move!
I realize this newsletter is mostly about generative AI, but that’s ok! First, the majority of our generative AI tooling is in fact, serverless. But second, it’s impossible to ignore the clear shift in software development trending toward the use of AI. From all aspects of development, we’re seeing fundamental changes. I’m not going to suggest we all fully embrace it and toss out everything we know, but we can’t ignore it - it won’t be going away.
If you’d like to make a recommendation for the serverless superhero or for an article you found especially useful, send me a message on Twitter, LinkedIn, or email.
Happy coding!
Allen
Thank you for subscribing!
View past issues.