In Season 5, Episode 12, Karl and Jon are joined by Farah Abdirahman, an AWS Community Builder, to discuss Amazon S3’s new account regional namespaces for general purpose buckets, deploying AWS applications and accessing AWS accounts across multiple regions with IAM Identity Center, AWS and NVIDIA deepening their strategic collaboration to accelerate AI, celebrating 20 years of Amazon S3, and Microsoft reportedly considering legal action over the recent $50 billion Amazon-OpenAI cloud deal. Then, just when things couldn’t get any more unexpected, the conversation took a turn toward the smell of Jon’s feet — and let’s just say the guys really put their foot in it.

07:28 - Amazon S3 Account Regional Namespaces

This feature allows S3 bucket names to be unique within an account and region, rather than globally. This change simplifies bucket naming conventions and addresses long-standing challenges with global uniqueness requirements. The impact is significant for daily operations and resource management in S3.

13:59 - AWS IAM Identity Center Multi-Region Deployment

AWS now offers multi-regional replication for IAM Identity Center, enabling users to access applications and accounts across multiple regions. This feature enhances resilience and reduces the need for break-glass setups. It also supports integration with external identity providers like Okta and Microsoft Entra ID.

21:05 - AWS-Nvidia AI Collaboration

AWS plans to deploy at least a million Nvidia chips in their regions this year to accelerate AI deployment. This partnership raises questions about AWS's own chip development efforts and highlights the increasing demand for AI-capable hardware. The collaboration also includes expanded support for Nvidia Nemo models on Amazon Bedrock.

26:25 - Amazon S3 20th Anniversary

S3 celebrated its 20th anniversary, highlighting impressive statistics such as 500+ trillion objects stored, 11 nines of durability, and 200+ million requests per second. The service continues to evolve and remains a cornerstone of AWS's offerings, with new features and improvements still being developed.

37:04 - Microsoft-Amazon-OpenAI Legal Dispute

Microsoft is considering legal action over the recent $50 billion deal between Amazon and OpenAI. The dispute centers on whether OpenAI can offer certain services without violating its previous agreement with Microsoft. This situation highlights the intense competition and large sums of money involved in the AI industry.

In Season 5, Episode 11, Karl and Jon are joined by Dmytro Sirant, AWS Community Builder and User Group Leader from Australia, to discuss the expansion of AWS Database Savings Plans, AWS European Sovereign Cloud compliance milestones, managing Amazon Machine Image lifecycles with AMI lineage, SpaceX’s plan for a million-satellite data center and Amazon’s opposition, and AI coding assistants and their potential impact on Amazon outages, with a few unexpected tangents along the way, including turtles and frozen corpses.

09:31 - AWS Database Savings Plans expansion

AWS has added Amazon OpenSearch Service and Amazon Neptune Analytics to the Database Savings Plans. This expansion provides more flexibility for clients who haven't decided which database best fits their requirements. The plans currently offer only one-year, no-upfront options, which is more limited compared to Reserved Instances.

16:48 - AWS European Sovereign Cloud compliance milestones

The European Sovereign Cloud has achieved its first compliance milestones, including SOC 2 and C5 reports, plus seven ISO certifications. These certifications are crucial for organizations requiring compliance and demonstrate that the European Sovereign Cloud is operating independently from AWS proper.

27:50 - Managing AMI lifecycles using AMI lineage

AWS introduced AMI lineage, a tool for managing the lifecycle of Amazon Machine Images. This solution helps track the chain of custody for AMIs, which is particularly useful in large enterprises with multiple teams working on image creation. However, it requires manual deployment and may be unnecessarily complex for many users.

32:35 - SpaceX's million-satellite data center plan and Amazon's opposition

SpaceX has filed plans with the FCC for a million-satellite data center in low Earth orbit. Amazon has objected to this plan, claiming it lacks substance and is purely aspirational. The discussion touched on potential issues such as cooling systems for satellites and the impact on astronomy.

40:08 - AI coding assistants and Amazon's outages

Recent outages on Amazon.com have led to speculation about the rapid adoption of AI coding assistants potentially causing issues. The discussion focused on the challenges of integrating AI-generated code into existing development processes and the need for improved review mechanisms to handle the increased output from AI assistants.

In Season 5, Episode 10, Karl and Jon discuss several developments in the AWS and cloud ecosystem, including the new output formats in AWS CLI v2 and how they improve usability and automation. They also explore the strategic partnership between OpenAI and Amazon and what it could mean for AI infrastructure and the broader cloud landscape. The conversation dives into architectural design as well, looking at rewriting Step Functions as Durable Functions in a Lambda-heavy approach, and how teams can use the AWS Well-Architected Framework to uncover hidden costs in their environments. They also touch on reports of AWS data centers in the UAE being targeted by Iranian drones, discussing the implications for cloud resilience and global infrastructure. And in a lighter moment, the guys compare notes on who drove the furthest for their hobbies last weekend.

02:30 - New output formats in AWS CLI v2

AWS has introduced new output formats in CLI v2, including an enhanced format for better error messaging and debugging. The update allows for suppression of CLI output, which is useful for handling sensitive information. These changes aim to improve user experience and security when working with the AWS CLI.

08:48 - Strategic partnership between OpenAI and Amazon

OpenAI and Amazon announced a strategic partnership where OpenAI will consume 2 gigawatts of Trainium capacity through AWS infrastructure. This deal involves significant investment from Amazon and allows for distribution of OpenAI's models via AWS Bedrock. The partnership raises questions about the economics and future of AI adoption.

18:14 - Rewriting Step Functions as Durable Functions (Lambda Heavy)

Danielle Heberling wrote an article about rewriting her Step Function as a Durable Function (Lambda Heavy). The post compares the two approaches, highlighting the benefits of Durable Functions for developers who prefer standard programming languages and fine-grained control over execution state in code.

28:53 - Using the AWS Well-Architected Framework to uncover hidden costs

The article discusses how the AWS Well-Architected Framework can be used to uncover hidden costs in cloud architectures. It emphasizes that hidden costs are not just about direct expenses but also include potential costs related to security breaches, downtime, and regulatory compliance.

34:58 - AWS data centers in UAE targeted by Iranian drones

AWS data centers in the UAE were targeted by Iranian drones, causing power outages and downtime for some applications. This event marks the first time data centers have been specifically targeted in a conflict, highlighting the need for multi-region resilience and raising questions about the future security measures needed for data centers in conflict zones.

In Season 5, Episode 9, Karl and Jon are joined by Jenn Bergstrom, VP of Cloud and Data Solutions at Parsons, AWS Community Builder, and AWS Ambassador. They discuss AWS Landing Zone Accelerator and custom CI/CD pipelines, AWS Security Agent’s new support for penetration tests on shared VPCs, A recent AI-augmented campaign that targeted over 640 Fortinet firewalls, Resilience testing on Amazon ElastiCache with AWS Fault Injection Service, AWS IAM Policy Autopilot now available as a Kiro Power, Controversy over whether AI (Kiro) or human error caused a recent AWS outage, and the conversation moved on to whether it’s more important to protect the feelings of human team members or AI agents!

04:48 - AWS Landing Zone Accelerator and custom CI/CD pipelines

The hosts and guest discussed the complexity of the new feature, comparing it to existing solutions like AWS Control Tower for Terraform. They questioned the necessity of the additional steps involved and wondered about its potential adoption rate.

13:15 - AWS Security Agent's support for penetration tests on shared VPCs

The speakers highlighted the importance of this feature for organizations with multiple accounts and shared VPCs. They discussed how it could complement traditional penetration testing services by allowing more frequent and cost-effective internal testing.

19:16 - AI-augmented campaign targeting Fortigate firewalls

The discussion centered on the use of AI tools by malicious actors to create more sophisticated attacks. The speakers emphasized the importance of basic security hygiene, such as changing default credentials and keeping management interfaces off the internet.

27:24 - Resilience testing on Amazon ElastiCache with AWS Fault Injection Service

A quick walkthrough for chaos-testing Amazon ElastiCache resilience with AWS Fault Injection Service, simulating an Availability Zone interruption/failover and showing how to run the experiment, capture logs, and validate your app’s reconnect/retry and fallback behavior.

36:18 - AWS IAM Policy Autopilot as a Kiro Power

The hosts and guest talked about the benefits of integrating IAM Policy Autopilot into Kiro, making it easier for developers to create secure IAM policies. They discussed the concept of Kiro Powers and how they simplify access to various AWS tools and services.

41:17 - Controversy over AI vs. human error in AWS outage

The speakers discussed the recent controversy surrounding whether an AWS outage was caused by Kiro (AI) or human error. They debated AWS's decision to blame human error rather than their AI tool, considering the implications for trust in both AI systems and human engineers.

In Season 5, Episode 8, Karl and Jon are joined by Jack McGuire*, an AWS Community Builder and software engineer at Green Man Gaming. They discuss how Amazon RDS now supports backup configuration when restoring snapshots, default encryption at rest for new Amazon Aurora clusters, AWS’s plans for space-based data centers and satellite constellations, Amazon’s Leo (formerly Project Kuiper) satellite internet service, and Amazon’s $200 billion Capex plan for AI data center buildouts, before the conversation once again turns to plumbing hacks.

04:27 - Amazon RDS backup configuration

This minor feature update allows users to change certain backup parameters when restoring RDS and Aurora snapshots, rather than having to modify them after restoration. While a small quality-of-life improvement, it's mainly beneficial for those managing instances through the console.

08:19 - Default encryption for Amazon Aurora

Amazon Aurora clusters now have encryption enabled by default, following the trend of other AWS services. This change simplifies security best practices and eliminates the need for manual encryption setup, particularly beneficial for new deployments.

14:10 - AWS space-based data centers

StarCloud is planning to launch satellites with AWS Outposts hardware, aiming to enable high-performance computing in space. The company intends to deploy up to 88,000 satellites, raising questions about the practicality and purpose of such a massive space-based infrastructure.

23:40 - Amazon Leo satellite internet

Amazon's low Earth orbit satellite internet service (formerly Project Kuiper) hints at opening its technology to third-party antennas. This approach differs from competitors like Starlink, potentially allowing for a more open ecosystem and wider adoption, especially in developing countries or remote areas.

34:51 - Amazon's $200 billion Capex plan

Amazon announced a significant investment in AI data center buildouts. While the scale of investment raised concerns, industry analyst Corey Quinn suggests that Amazon can likely handle potential market fluctuations better than smaller competitors due to its size and resources.

*More about our guest Jack McGuire:

In Season 5, Episode 7, Karl and Jon are joined by - Kai (Huynh Le Nhat Nghia), AWS Community Builder from Vietnam. They discuss cross-account stream processing with AWS Lambda and DynamoDB - choosing between Amazon ECS Blue/Green native or AWS CodeDeploy for deployments - how AWS Support helped refund unexpected QuickSight charges - Amazon's $200 billion investment in AI data center capacity - growth of sovereign cloud spending in Europe - and once again the guys went off on a tangent about Jon's feet...

04:06 - Cross-account stream processing with AWS Lambda and DynamoDB

This article discusses a new feature allowing simplified cross-account stream processing. It highlights use cases such as centralized data processing, shared services, and multi-tenant architectures. The feature provides improved security by maintaining data ownership in one place while allowing controlled access from other accounts.

10:13 - Choosing between Amazon ECS Blue/Green native or AWS CodeDeploy

The article compares ECS native blue/green deployments with using AWS CodeDeploy. While ECS native offers simpler deployments, CodeDeploy provides more advanced features like canary deployments, bake periods, and lifecycle hooks. The choice depends on the complexity of deployment requirements.

19:44 - AWS Support refunding QuickSight charges

A user shared their experience of unexpected QuickSight charges after a promotional period ended. AWS Support helped refund the charges. The article emphasizes the importance of setting budget alerts and understanding service pricing, especially after promotional periods.

27:00 - Amazon's $200 billion AI data center investment

Amazon announced a $200 billion investment in data center capacity for AI, with plans to double capacity by 2027. The podcast discussion touched on whether this massive investment is part of an AI bubble and how it might be monetized in the future.

34:16 - Growth of sovereign cloud spending in Europe

The article discusses the expected tripling of sovereign cloud spending in Europe, driven by concerns over data sovereignty and the US Cloud Act. The podcast touched on the challenges of truly sovereign clouds when subsidiaries are still owned by US companies.

In Season 5, Episode 6, Karl and Jon are joined by Danielle Heberling, an AWS Hero, to dig into AWS Serverless updates from Q4 2025, including Amazon DynamoDB Global Tables now supporting cross-account replication, a new feature for displaying related resources for EC2 and VPC security groups, and IAM Identity Center now supporting multi-region replication. They also explore the feasibility of orbital data centers—before the conversation takes a fun turn as the guys debate whether data centers in space would technically be “in the cloud” or above it.

06:14 - AWS Serverless updates from Q4 2023

This article summarized various serverless updates from AWS in Q4 2023, including Lambda durable functions, Lambda managed instances, provision mode for SQS, ECS express mode, and other improvements. The speakers discussed the benefits of these updates, particularly highlighting durable functions as a significant quality of life improvement for developers.

14:53 - Amazon DynamoDB Global Tables cross-account replication

AWS announced support for replicating DynamoDB Global Tables across different AWS accounts. The speakers discussed potential use cases, such as disaster recovery and data lake setups, but also questioned the necessity of this feature given existing workarounds.

21:31 - EC2 and VPC security group related resources display

A new feature was introduced allowing users to view related resources for EC2 and VPC security groups in the AWS console. The speakers agreed this would be beneficial for operators, making it easier to understand and manage security group configurations across various services.

25:03 - IAM Identity Center multi-region replication

AWS announced support for multi-region replication of IAM Identity Center configurations. The speakers discussed the potential benefits for disaster recovery and compliance scenarios, as well as how it might simplify break-glass access procedures.

33:04 - Orbital data centers

The podcast concluded with a discussion on the feasibility of orbital data centers, prompted by an article where AWS's CEO called them "pretty far from reality." The speakers debated the technical challenges, costs, and potential benefits of space-based data centers, ultimately agreeing that while interesting, such technology is likely far off in the future.