AWS Bites Podcast

106. Luciano's reInvent Experience

In this episode, Luciano and Eoin chat about Luciano's experience attending AWS re:Invent 2023 in Las Vegas for the first time. They talk about the massive scale of the event, logistical challenges getting around between venues, highlights from the keynotes and announcements, and tips for networking and getting the most out of re:Invent. Luciano shares his perspective on the AI focus, meeting people in real life after connecting online, rookie mistakes to avoid, and why re:Invent is worth the investment for anyone working in the AWS space.

AWS Bites is brought to you by fourTheorem, the ultimate AWS partner for modern applications on AWS. We can help you to be successful with AWS! Check us out at fourtheorem.com!

In this episode, we mentioned the following resources.

Let's talk!

Do you agree with our opinions? Do you have interesting AWS questions you'd like us to chat about? Leave a comment on YouTube or connect with us on:

Luciano
00:02:36

Yeah probably made many at this point but yeah the main one is probably trying to plan to attend talks but not realize again how far away things can be. For instance yesterday I wanted to attend the power tools workshop and I had to go from the Venetian which is where I'm staying to Mandalay Bay which if you just look on Google Maps doesn't seem too far away and there are buses so I left like one hour early with the bus from the Venetian and it got stuck in traffic and it took slightly more than one hour so when I got there and I found the room it was probably like five ten minutes late so not too too late but the room was already packed and they just didn't let me in so it was a bit of a shame that I spent like more than one hour trying to get there to attend the workshop and then I couldn't attend it. I was still able to meet the power tools team and have a nice chat after it so it was totally worth it to go anyway but yeah it was a bit of a rookie mistake I should have planned that a little bit better and maybe go there a little bit earlier so I think that kind of stuff can happen and sometimes there are so many talks so you might not realize when you create your own agenda that maybe one after the other is just not feasible because you're not going to have enough time to to move from one place to the next one so someday definitely to keep in mind and next year maybe try to have just a very few talks you want to attend and make sure you have enough space to move from wherever you are to the to the venue where they're doing the talk.

00:06:09

And yeah it was actually a really good talk by Efi and he is also another serverless hero and basically he actually described three different ways to use rust in the context of Lambda so not just building on Lambda runtimes but even just packaging a library written in rust for python for instance and then just using it in the python runtime or even building custom Lambda extensions with rust and then using them with other runtimes like java or not js so a bunch of interesting learnings there and the conversation after it was super interesting it was actually relatively small crowd which wasn't surprising because it's a very niche topic still but everyone that was there was like super excited and everyone was trying to share all the mistakes and all the lesson learned while trying to do this thing which is still pretty new and not many people are doing so it was quite an engaging session because everyone was super excited to try to learn from each other and other than that has been mostly trying to meet people I wasn't really sure how to schedule all of that again because my first time here I didn't know what the spaces will look like what everyone agenda is going to look like so I was just connecting before the event online on slack linkedin twitter and just asking people I'm going to be there let's exchange maybe phone numbers and then let's try to find some time to spend some time together and that has been working out in in a mixed way I think in some cases it's working very well in other cases I just was unable to meet the people because the schedules didn't really match but also there have been a few interesting events organized by AWS for the heroes or the community builders so being able to to attend this focused events is generally the best place to meet people that more or less have the same interest and people that you might want to talk to and just share opinions.

Eoin
00:10:31

There are a couple of nice ones in the kind of serverless space like SQS FIFO throughput has had a massive increase and now you can do 70,000 transactions per second so now SQS FIFO queues used to be a lot more limited in terms of throughput now with 70,000 transactions per second you can theoretically put 700,000 messages through it in a second and also with SQS FIFO you can now do QA drives for DLQs which was something that was missing before and then on Lambda the scalability of Lambda has suddenly had another jump as well so everybody can get a thousand invocations per minute after the burst concurrency we talked a lot about burst concurrency and Lambda in a previous episode but like in most accounts that we've used in eu-west-1, we can get 3,000 immediately and then it used to be 500 per minute. Now, that's a thousand per minute, so you can scale up really fast. And then on Step Functions, there were two announcements that I'm aware of. One was the ability to be able to redrive from a failed state. I think that's a really good one so now you don't have to run the whole function again you can just continue from the failed step. And the other thing is something we actually asked for I think when we covered Application Composer which is the ability to integrate Step Functions workflow studio into Application Composer now that's done so they're obviously listening thanks everybody what I think you caught a lot of the Adam's Selipsky keynote yesterday now there's a lot of AI as we could have predicted and I think even though they're pretty impressive in a lot of ways all of the Amazon Q announcements we could have predicted them as well because it felt like AWS had to have an answer to all the announcements from the open AI developer conference from a few weeks ago do you have any thoughts I'm guessing you haven't had a chance to try it out but have you heard from anyone is there a mood in Las Vegas on what Amazon Q can do? Is it skeptical or optimistic?

Eoin
00:13:27

I've had a very brief period just playing with the Q assistant in the AWS console because it's like a it's almost a bit like Clippy for AWS console it's like, "I see you're starting an EC2 instance. I can help you with that!" I've got mixed results so far but you know we know AWS tends to launch things a little bit early especially like they were it seemed like they were going to have to launch this no matter what I would give it time I'm sure it will improve models will improve anyway but I actually am impressed by the way they've integrated Q into AWS console visual studio code AWS toolkit loads of other services like connect and quick site it must have been a massive massive undertaking to get everything ready in time for re-invent so that is impressive and there's there's loads more like you can you can start your own chatbot application you can integrate it with your own data sources like Gmail and Dropbox and Aurora, SQL server loads of stuff and it's it's an alternative which I would probably intuitively trust a lot more than any of the other offerings out there because I know about AWS security and how it works and I would have high confidence in it doing the right thing with my data I'd be much more comfortable building like an enterprise knowledge base retrieval augmented system with a bedrock or Q than I would with anything else and there's obviously a lot of focus on code transformation as well they announced this Q code transformation that can migrate Java 8 applications to Java 17 applications I think in minutes now the Adam's claims were quite extreme I think it was something like thousands of applications migrated in minutes I'd love to know how that works and how do you even test them exactly I don't mean that amount of time but let's see how it goes. I've started playing around a little bit with the AWS toolkit in VS code so loads of stuff to play with there I think it's worth giving a try and I'm sure there'll be some rough edges at the start but as alway.s it'll improve over time so apart from announcements Maybe some advice, what would you recommend based on what you've seen so far other people to go?

Help us to make this transcription better! If you find an error, please submit a PR with your corrections.