🦸 Community Superhero

Our community superhero this week is Linda Mohamed, AWS Community Hero, cloud consultant, and rock star engineer. If you’ve ever had the pleasure of meeting Linda or watching her talk, you know she’s a brilliant engineer with a knack for building with AI and believe it or not, juggling! Congratulations on becoming a lecturer for university students, Linda, and thank you for everything you do and sharing your contagious energy!

💯 Spotlight

Our spotlight this week is actually a post I missed from two weeks ago 😬. But it’s a good one from our trusted friend Danielle Heberling, so it’s worth surfacing again this week. Danielle lays a claim that not all AI builders are doing the same work and I agree. Her article focuses on the different types of AI used to do development work (no spoilers here) and I like the way she breaks it all down. As with everything, I think there’s some nuance in there and I could probably get nit picky, but it’s good to categorize the types of work into different buckets.

🔥 My Favorite Content

By now we’ve seen DynamoDB vectors a few times. But what we’ve seen is mostly theoretical and doesn’t leave you with a solid “what’s next?” Last week, we got our first steps from Andres Moreno with a walkthrough of how to add semantic search to an existing DynamoDB table. He shows you the code to enable it (because it’s not supported by CloudFormation yet) and how to build a basic search. He also gives you some handy things to consider before you start building.

And what good is agentic semantic search without memory? I published an article last week on agentic memory on Valkey, covering the lessons I learned recently while I was building a demo. Valkey is as fast as you’d think for memory retrieval, but I discovered there are some important things you need to know up front before you start building. I cover the two big ones that bit me in the post to hopefully save you some time and effort.

For a more concrete break down on memory compared to what I wrote about, I highly recommend Ken Alger’s article from last week. He takes some of the lessons I learned and runs with them. He goes into the whole science behind memories and memories superseding each other and how in general memory is not a solved problem. I really like the depth of this article and it makes you think twice about simply adding on a memory object through your SDK to bring stuff up later.

💡 Tip of the Week

A revolution isn’t a revolution if you simply swap something out for another. Replacing people with AI agents and asking them to do the same thing doesn’t move us forward. Quite the opposite. Emiliano Rodriguez made a good point last week and even wrote a short article on the subject that’s worth the read.

Last Words

There’s lots going on in the world of AI right now (obviously). I find it so fascinating to watch the focus areas pop up week over week, like agent memory and search was in this issue. I wonder what the next one will be?

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