I read the viral "leaving AWS" post twice.


AWS FOR THE REAL WORLD
⏱️
Reading time: 12 minutes
🎯
Main Learning: Most of the complaints in the viral "leaving AWS" post are skill issues β€” but egress pricing is a fair hit.
πŸ“

Hey Reader πŸ‘‹πŸ½
Recently, a post with the title "I returned to AWS and was reminded why I left" hit 810 upvotes on Hacker News last week and went pretty viral with it.
I read it twice before forming an opinion. My honest take: most of the complaints are skill issues! πŸ€·β€β™‚οΈ
Nevertheless, the post is well written and I really couldn't stop myself from writing a detailed response about why most of it reads like someone used AWS the "wrong way" and blamed AWS for the outcome.
In this issue, I go through all the complaints one by one. Most things I disagree with while one is definitely legit.

At the end I explain the pattern I see across all of them 🫑

Sponsored by Spacelift
Free Guide

Find your Terraform maturity stage in 30 minutes.

Spacelift's free guide walks platform teams through the three maturity stages of running Terraform at scale. Spot which one you're stuck in.

Terraform State at Scale: The Three-Stage Maturity Model for Platform Teams

Spacelift is an infrastructure orchestration platform for Terraform, OpenTofu, Pulumi, CloudFormation, Ansible, and Kubernetes. Used in production at Duolingo, Figma, and 1Password.

This issue is sponsored by Spacelift.

A massive sledgehammer hitting a tiny thumbtack

πŸ“š This Week's Deep Dive

A post about leaving AWS hit 810 upvotes on Hacker News last week. The frustration in it is real and the writing is sharp. But most of the ten complaints disappear the moment you look at how the platform is actually supposed to be used.

A few of the author's takes and the short version of why I disagree:

On IAM

"IAM β€” the hideously complex auth and access rules system. This was invented by Lucifer sitting on his burning throne in the ninth level of Hell as the worst possible torment for those who have been sent below for using AWS."

The core model is three questions: who, what action, on which resource. Lucifer's off the hook.

On DynamoDB

"DynamoDB what a hot pile of garbage. I tried it and ended up with a $75USD bill by the end of the day."

I ran DynamoDB for a million users on a few hundred a month. The difference between that and $75-in-a-day is whether you run full-table Scans.

On Lambda

"There's simply no genuine benefit to AWS Lambda compared to running your own web servers… Keep convincing yourself that using AWS Lambda is not a horrible mistake."

On event-driven workloads, "horrible mistake" is the wrong framing. Cold starts are well under 0.1% of invocations in production.

πŸ“° This Week in AWS

πŸš€OpenSearch Serverless next-gen is GA

Scale-to-zero pricing replaces the old cluster-provisioning model, with up to 60% savings on spiky workloads and 20x faster scaling. The first time OpenSearch fits agentic-AI traffic patterns instead of fighting them. Read More β†’

πŸ”§Bedrock exposes mantle quotas through Service Quotas

The bedrock-mantle endpoint (where OpenAI and Anthropic APIs run on Bedrock) now reports per-model token quotas via Service Quotas. Same surface as bedrock-runtime, so you can spot rate limits before prod traffic hits them. Read More β†’

That's it for this issue!
If you take one thing away: use the right tool for the right job. AWS is built for production at scale through IaC and you should know what you're doing. It's not the best "weekend hobby project platform" - we can be honest about that 😬

​
See you in the next one!

Sandro & Tobi

AWS for the Real World

We teach AWS for the real world - not for certifications. Join more than 10,500 developers learning how to build real-world applications on AWS.

Read more from AWS for the Real World

⏱️ Reading time: 15 minutes 🎯 Main Learning: A public container on AWS for $7 a month, with no VPC to own. πŸ“ Blog Post 🀝 Sponsored by AWS. What we built, what it cost and what we think of it are our own. Hey Reader πŸ‘‹πŸ½ Your first container on ECS comes with homework: a VPC, subnets in two Availability Zones, a load balancer, security groups, a certificate and a task role. Express Mode creates most of it for you now (App Runner's successor, so ECS abstraction number three), but it all lands in...

AWS FOR THE REAL WORLD ⏱️ Reading time: 6 minutes 🎯 Main Learning: You cannot skip a level of AWS observability. Structured logs come first, then the AI agent that reads them for you. πŸ“ Blog Post πŸ“‚ Claude Code agent (free gist) 🎬 Watch on YouTube Hey Reader πŸ‘‹πŸ½ This week I'm bringing back one of our most-read posts: the 3 levels of AWS observability. A lot of you joined after it went out in May, and the topic fits this week perfectly. Because the real question isn't "do I have logs?" anymore....

⏱️ Reading time: 11 minutes 🎯 Main Learning: Measure NAT traffic first. CI runners rarely need NAT at all. πŸ“ Blog Post Hey Reader πŸ‘‹πŸ½ NAT Gateways are one of those things you set up once and never look at again (probably!) πŸ˜… Until NatGateway-Bytes shows up near the top of the bill, filed under EC2-Other - thanks AWS! I think they've made this on purpose to hide these significant networking related costs somewhere and made it hard to reflect on! The hourly fee is predictable, sure. But the...