Edmund Haselwanter

On my way to the AWS Summit in Berlin

My #does23 talk on "AI, Make it So - a dive into prompt engineering" video is now available ✨ - I explain how our daily engineering life is changing with these new technologies . Enjoy 20minutes of goodness - RT = ❤️ youtu.be/EDDa5bu480Q

So my question about OSS business models triggered quite a number of tweets @adamhjk ;) :). Thx for the great insights!

I love how training llms for code generation thrive on code Examples , documented functions and tests. So, if not for a collegue , do it for the AI so it learns better and helps you more :)

Checkout winglang.io if you haven't. That really looks awesome @emeshbi

Awesome talk by @adamhjk about technology as art #cfgmgmtcamp, and how it inspired @thesysteminit. I was too busy listening so I didn't have a chance to take any pics. Definitely watch it later!

And now @patrickdebois is kicking it off

Arrived in Ghent for Cfgmanagementcamp cfp.cfgmgmtcamp.org/2024/schedule/

1. It's config management time, so here is a thread about configuration languages in IT. Config languages are not unique to IT, they exist in all areas of science and engineering to describe layout or "state".

@adamhjk See you there :)

Writing this talk for #cfgmgmtcamp, and it's a hard one. Trying to figure out how to write survey's of the technology landscape, keep things upbeat, and still have an opinion.

@patrickdebois See you there :)

See you in Ghent , right after Fosdem ? I'll be talking on the collaborative intelligence with GenAI agents (not yet sure what that even means yet :)

I was asked how long conversational programming will take to become the new "norm". I published this back in 2019, it's still pretty good as far as I'm concerned.

📢 KubeCon North America 2023 was a blast, and the observability track was particularly fascinating. 🚀 The Tracetest team had a booth and attended the co-located Observability Day event. Read our recap below! #KubeCon #Observability #OpenTelemetry tracetest.io/blog/observabi…

@mitchellh Have you tried out Cool Retro Terminal github.com/Swordfish90/co…

Have you ever wanted your terminal to look like an authentic CRT? Of course you have. Well, LOOK NO FURTHER because the future (or is it the past?) is now. And people ask why you'd ever need a GPU-rendered terminal.... HAH! My yak has no more hair to shave. This demo is crazier than it seems: you can actually specify any custom shader, and the shader API is compatible with ShaderToy, so you can take any off-the-shelf shader from ShaderToy and this works. The tech behind this is super over the top. Ghostty is cross-platform and uses Metal directly on macOS and OpenGL on Linux. Metal and OpenGL use different shading languages. No problem! You specify the custom shader in GLSL (OpenGL language) and it’s automatically translated to SPIR-V (binary IR) then automatically translated to MSL (Metal language). That all happens at runtime because why not. Yeah but this has to kill your battery right? Nope! <1% CPU and <1% GPU because computers are fast (especially when you're working at a relatively low level).

github.com/fr0gger/Awesom… #devsecops

Lessons learned on bringing GenAI application towards prodction - prompts, models, security , testing , pipeline , observability, cost, pricing ... Extended version from All Day Devops talk - Subscribe to my channel for more , feedback is ❤️ RT = Karma youtu.be/7Qr1f4f8u64

Discover how HashiCorp Vault seamlessly melds with ACME and Terraform configurations, improving your security. Find out about innovative approaches to certificate integration and boost your tech expertise. Read more: buff.ly/3QmE7ub #TechIntegration #hashicorp

OpenTelemetry removes the need to have multiple instrumentation libraries and streamlines data collection across your observability backend. thenewstack.io/what-is-opente… #Observability #OpenSource @opentelemetry @chronosphereio

Introducing MemGPT 📚🦙 a method for extending LLM context windows. Inspired by OS mem management, it provides an infinite virtualized context for fixed-context LLMs. Enables perpetual chatbots & large doc QA. 🧵1/n Paper: arxiv.org/abs/2310.08560 GitHub: github.com/cpacker/memgpt