AWS Should Stop Playing the Catch Up Game | ✉️ #45
Hey! 👋
Adam Selipsky was the CEO of AWS for the last 3 years. Recently, the company announced that he is stepping down and will be replaced by Matt Garman, someone who was growing inside AWS for the last 18 years, starting initially as an intern. AWS’s growth declines, and its market share shrinks as of today. I guess, it’s only natural to search for the new head of the business to steer it into the right direction. What this direction might be?
Recently, it feels like AWS is not as customer-obsessed as it used to be. It tries to catch up with the AI hype train, invests billions into its AI offering and doesn’t seem to care much about anything else. I mean, of course they still have thousands of engineering who continue building and improving existing services, but you can clearly see where the focus is.
The problem with this focus on AI is that there are thousands of businesses, or perhaps, the majority of businesses that don’t really need GenAI inside their products - unless they want to follow the hype train and introduce an LLM-powered Chat Bot. But what do they need? That’s the question that AWS seems to ask less and less often, leaving many of the recently released services, like AppRunner or Lattice, in a semi-abandoned state.
I’d love to see Matt Garman bringing the focus back to doing what’s best for the customers instead. Of course, they need to catch up with Azure and Google Cloud in AI space. But they need to remember, what made them the market leader in the first place - and that was not playing a catch up game.
What We've Shared
StackOverflow and OpenAI, Spotify BackStage Portal, Oracle23ai GA, AWS Bedrock Studio / mknews #5
EU Sovereign Cloud, New AWS CEO, Cloud Lock-In and other news from AWS Summit Berlin 2024
DevOps Accents #37: Broadcom and VMware, Stack Overflow and OpenAI. Will the VMware acquisition affect your company? Is there a problem with Stack Overflow partnering with OpenAI? Join Leo, Pablo and Kirill in a discussion of all the recent DevOps, Cloud and AI controversies.
And on the website please welcome the first of two parts of a big overview by mkdev's Head of Data & AI, Paul Larsen:
Which database when for AI: Are vector databases all you need? Paul explores semantic search with Generative AI and LLMs, discussing Retrieval-Augmented Generation (RAG). He compares vector and relational databases using German Masters Swimming data and provides insights on selecting the right database for business needs and AI strategy.
All while Pablo explains how to upgrade Openshift 4.x:
What We've Discovered
Introducing cloud console cartographer: It's an Open Source tool that helps security teams to track end to end what happened in AWS, based on CloudTrail events.
Minimizing on-call burnout through alerts observability: CloudFlare's observability team shares details about the company's Prometheus and Alertmanager setup. A lot of interesting technical details, though not as much info on reducing alert fatigue as the beginning of the article assumed.
Improve cost visibility of Amazon EKS with AWS Split Cost Allocation Data: AWS now allows you to breakdown EKS costs down to the pod level, and even tie the costs to CPU and Memory utilization (as long as you are using Manage Prometheus service). This is a great addition to AWS FinOps stack, as most EKS clusters out there are either multi-tenant or multi-application systems, and granular cost data is very much needed.
Solving Observability's Cardinality Conundrum: An exploration of the influence of cardinality on observability, in particular on how it affects the metrics.
Introducing Defense against AI-guided Traffic Analysis (DAITA): Hackers now can read traffic with help of AI analysis and there are tools that help to prevent this.
An announcement
Do you want to meet the mkdev team this summer? Check out our Event Schedule for events happening in Madrid, Munich and Belgrade!
The 46th mkdev dispatch will arrive on Friday, June 6th. See you next time!