AWS re:Invent 2023 conference is setting off just now, and NordHero has again sent three cloud architects to Las Vegas to find out the new and noteworthy on everything AWS and the whole ecosystem has on the plate. We also put up our wish list on new services, new features and the direction of development we would like to see during re:Invent 2023.
The week at the end of November is the most important week of the year for us at NordHero. Although the autumn season has been one colossal pre:Invent with dozens of excellent new features and service announcements, the biggest and boldest ones have been reserved for this week. NordHero has again sent three brave Cloud Architects, Janne, Ville, and Niklas, to Las Vegas to hear the news from the source and connect with AWS product teams and peers.
But what do our Cloud Architects wish from this year’s announcements?
Outpouring of Generative AI services and integrations
Earlier, we had AWS SageMaker Foundational Models. Then came Amazon Bedrock. The first offered an excellent way of getting your hands on Generative AI services through AWS SageMaker. The second offered an easy API for integrating with many GenAI services available through Amazon Bedrock and securely and privately customizing the Foundational Models with your organization’s data to meet your requirements better.
What we expect from AWS re:Invent 2023 is a series of ready-made integrations of Amazon Bedrock powered GenAI services with other AWS services. And during the last 24 hours, we just heard of the first ones, so here we go!
- AWS Step Functions launches optimized integration for Amazon Bedrock: https://aws.amazon.com/about-aws/whats-new/2023/11/aws-step-functions-optimized-integration-bedrock/
- AWS Config launches generative AI-powered natural language querying (Preview): https://aws.amazon.com/about-aws/whats-new/2023/11/aws-config-generative-ai-powered-natural-language-querying-preview/
- Powered by foundation model, Amazon Transcribe now supports over 100 languages: https://aws.amazon.com/about-aws/whats-new/2023/11/amazon-transcribe-over-100-languages/
- Descriptive Bot Builder with Generative AI: https://aws.amazon.com/about-aws/whats-new/2023/11/descriptive-bot-builder-generative-ai/
- Amazon Detective introduces finding group summaries using generative AI: https://aws.amazon.com/about-aws/whats-new/2023/11/amazon-detective-group-summaries-generative-ai/
- Announcing AWS Console-to-Code (Preview) to generate code for console actions: https://aws.amazon.com/about-aws/whats-new/2023/11/aws-console-to-code-preview-generate-console-actions/
- Assisted Slot Resolution with Generative AI: https://aws.amazon.com/about-aws/whats-new/2023/11/assisted-slot-resolution-generative-ai/
Continuum for Zero ETL Integration announcements
Last year, AWS CEO Adam Selipsky announced the term Zero ETL Integration. That means that AWS is pushing to integrate its services so that customers don’t need to invest in interconnecting AWS services. We already heard of some of those new integrations during re:Invent 2022, and we expect to listen to a bunch more this year.
Speeding up CDK
What CDK does under the hood is that it transforms your code into CloudFormation stack sets and deploys your infrastructure using CloudFormation service. If you use CDK to deploy more extensive infrastructures with hundreds of configurations, generating CloudFormation and especially running the CloudFormation tends to get very slow. And if you need to code and test the deployments several times per day, or if you need to push your code across many AWS environments with automated deployment pipelines, you spend a lot of your day just waiting for CloudFormation to deploy your code.
So, our wish for this year’s re:Invent is that the time-waste problem with CDK-based deployments is solved. It would be nice if CloudFormation were announced to start getting benefits from the Lambda SnapStart, which was announced last year for Java Lambda runtimes. Moreover, our wish list has a dream that CDK would support using OpenTofu (community-led, open-source Terraform fork) as an alternative to CloudFormation for deployments. Boy, that would be AWSome!
Improvements for Python and Rust developers
Boto3 is a commonly used AWS SDK for Python. Even though Boto3 has a comprehensive set of AWS services supported, it lacks native support for asynchronous programming. There are ways around it with 3rd party libraries, but we would appreciate the release of Boto4, an AWS SDK for Python that would be async from the ground up. Read more on the topic of building concurrency with AWS Lambdas from Ville’s blog: https://www.nordhero.com/posts/aws-lambda-concurrency/
Also, we would be happy to see native support for Rust language with AWS services, as Rust is gaining space with developers and is a highly-performing programming language. Read more from Vesa’s blog: https://www.nordhero.com/posts/rust-vs-typescript-on-aws-lambda/
New Deep* toy
Last but not least! It’s not a wish list if you don’t have a toy in it. It’s been a while since AWS announced a Deep gadget. The latest was in re:Invent 2019 when they announced AWS DeepComposer, the world’s first musical keyboard powered by machine learning to enable developers to learn Generative AI while creating original music outputs.
So we wish it would be time to release a new Deep* toy, or should it nowadays be called GenAI-toy?
With this wish list, we hope everyone very re:Inventing conference week, and if you are onsite in Las Vegas, pop up to say hi to our Janne, Ville, and Niklas!