Every blog you write is feedback — to OSS maintainers, to product teams, to the community. Your opinion matters. Don't let the flood of AI-generated content stop you from sharing what you've learned. Speak up and never stop! What's stopping you from writing? Let me know 👇

This is exactly what I was talking about! Markus is back to writing after 10+ years. No SEO strategy, no editorial calendar — just thoughts shaped through conversation, published when ready. Welcome back to blogging! https://lckhd.eu/uP2Hek

I'm actually also going to host an #AWSCommunityBuilders meetup at the #AWSCommunityDayDACH26! I'd love to see you there if you're a community builder to answer all of the questions that you have! See you on the 15ths of september! 🧵

A few of you might be building with Claude today - Ran shared his learnings and best practices with me - do they match yours? https://lckhd.eu/QqapPU #awscommunity #claude #genai #ai

A few of you might be building with Claude today 🔥 Ran shared his learnings & best practices with me — do they match yours? 👇 https://lckhd.eu/pC1eJb #awscommunity #claude #genai #ai

At the AWS Heroes Summit we talked about how many of us have gone quiet in 2026. AI floods the internet with content — but YOUR voice, your experience, your mistakes... that's irreplaceable.

Tonight we are meeting at the #AWSUGBergstrasse to talk about using tokens from k8s! Join us if you can make it - would be great!

We're kicking off our UG meetup now! Looking forward to a fantastic evening! #AWSCommunity #AWSUGBergstrasse

We're up tonight for a great topic #AWSUGBergstrasse TODAY at 7pm CEST in Bensheim "You are Paying to Compute the Same Tokens Twice" 🧵

AI changes the way that I look at the speed of delivering features. We shouldn't talk about "daily" deployments but rather about hourly. The biggest challenge will be how to safely deploy, speeding up your pipelines and ensuring that everything you deploy is tested. #BuildInPublic #AI

AI makes it "too simple" to generate blog posts in 2026. But your own thoughts, your battle scars, your decisions — that's what makes content valuable. I wrote about why writing and sharing knowledge still matters: https://lckhd.eu/QYwLE0

I think I've been able to fix a few problems in the game and now I know its going to be fun! If you've ever played ogame or planetation, this might be for you.... #BuildInPublic

I think I've been able to fix a few problems in the game and now I know its going to be fun! If you've ever played ogame or planetation, this might be for you.... #BuildInPublic

Quadratic compute. On every single step. That's what LLM inference looks like without a KV cache. The KV cache fixes it — making inference linear & affordable. But your cluster might be destroying that cache without you even knowing. 😬 AWS User Group 🧵

One of the underrated videos on my YT is one I did with Ran around Claude Best practices and things you need to know - somehow it didnd't get much traction... Why do you think thats the case? #Claude #ClaudeBestPractices #AIEngineering

If you haven't yet, register to our AWS community day DACH on the 15ths of september in Berlin - I hope I'll see you there! #AWSCommunity Link to registration in the comment...

A great blog from Allen. Funny enough Nexus Share is implemented in exact the way he recommended 🙃❤️ https://lckhd.eu/5OcaLE

Hot take: Kubernetes was never built for LLM inference. 🔥 Every scheduling primitive you trust — round-robin, readiness probes, HPA, rolling updates — was designed to IGNORE the exact state that makes inference cheap. The result? Turn 10 of a conversation. 96,000 🧵

The #AWS foundation for my #BuildInPublic projects is "Serverless", here's some details of how I built on GitHub https://lckhd.eu/O28zKR

build-in-public/docs/aws-foundation.md at main · Lock128/build-in-public

Contribute to Lock128/build-in-public development by creating an account on GitHub.

lckhd.eu

Can we get a few more of you attending the #AWSUGBergstrasse TODAY at 7pm CEST in Bensheim? We got a great topic! RSVP here: https://lckhd.eu/tje2HI

You are Paying to Compute the Same Tokens Twice, Thu, Aug 27, 2026, 7:00 PM | Meetup

This time we have a new talk by Christopher Haar Every token an LLM generates attends to every token before it. Without a cache, that is quadratic work on every single ste

lckhd.eu

Context switching is a big issue when working with coding agents. You need to always keep in mind what you gave the agent. Difficult if you work on different projects and different tasks in parallel. How do you deal with that? #BuildInPublic #BuildWithAgents

#AWSHeroes like to talk and share their opinions. We have a few options to chat and discuss. What I enjoy most is learning from what others share and complain about and this is why #AI will not replace us ever.

Yes, if you can post things from the place that you're "right now" (ide, cli, etc.) it's more common to write shorter posts. This is potentially where X/Bluesky will get more traction from me right now? How do you interact with social media? #SocialMedia #MCP

There's too many platforms to launch your #BuildInPublic product or SaaS on Really hard to follow and decide which one to use. I've crafted a few details on GitHub - which ones do you use today? https://lckhd.eu/BpVboF

Now that my MCP server works (again) and has 35 tools available, I can use this to post more human content again - directly from #Kiro @kirodotdev Or maybe also from #KiroCrew in the future?

Do you still use an RSS reader? Here's my RSS feed for my blog: https://lckhd.eu/9rAxhG

🧵 Every token an LLM generates attends to every token before it. Without a KV cache? That's quadratic work — recomputed on every single step. With a KV cache? Linear. That's why inference is affordable at all. But here's the catch: that cache lives on ONE specific 🧵

📊 App Stats Update (last 30 days) 🚀 3,987 new installs across 8 apps 🔄 7,239 app updates 📱 1,776 active devices (Android) 🍎 ~26 daily active (iOS) • Gem,Ore & Struct Finder for MC: 1,738 installed, ~26/day iOS, +3,976 new, 7,213 updates • Nexus Share - Social 🧵

#BuildInPublic is what I do with a few things, I am sure you've read some of it. Some details on the frontend technologies I use are on GitHub - What do you use today? #Flutter #Angular #Typescript https://lckhd.eu/popKrS

build-in-public/docs/tools-for-building.md at main · Lock128/build-in-public

Contribute to Lock128/build-in-public development by creating an account on GitHub.

lckhd.eu

Thursday at the AWS UG Bergstrasse: "You are Paying to Compute the Same Tokens Twice" This time we have a new talk by Christopher Haar Every token an LLM generates attends to every token before it. Without a cache, that is quadratic work on every single step, the same 🧵