Hierarchy, consensus, adversarial debate, and reality show knock-outs. We can design more reliable multi-agent systems with inspiration from human systems but there are fundamental differences and workarounds. #ai #llm #systemDesign #reliabilityEngineering blog.alexewerlof.com/p/multi-agen...

Do LLMs have emergent properties or not? As an interactive systems engineer, I feel I have to set the record straight: emergence isn’t an objective phenomenon and due to the architecture of LLMs, they sure have emergent (both weak and strong emergence). blog.alexewerlof.com/p/emergent-p...

Google Antigravity is so new the Gemini extension running in the same window has no clue about it! 🤯 Upon initialization, Antigravity installed all my VS Code extensions, one of which is Gemini Code assist! See the conversation below!!!

✋ Do you want to run your own AI with zero programming effort? I've made a completely free and open source product called Local Browser AI for Google Chrome Built-in AI Hackathon. You can check it out and maybe vote for me, if you feel like it is worth it. devpost.com/software/loc... #llm #ai #oss

. @charity.wtf (Honeycomb) interviews @alexewerlof.com(Volvo Cars) on the gap between Google's SRE ideals & reality. Key insight: SLOs are your team's API for pushing back against micromanagement.

This didn't age well #gpt5 #openai #samAltman

Enjoyed the different examples of SLIs types modelled here by @alexewerlof.com #SRE #SLI blog.alexewerlof.com/p/sli-compass

How do you assess the quality of a metric? How do you balance the cost measurement against ROI and complexity? #SLI Compass is a 2D mental model to facilitate evaluation, investment, and evolution of metrics. Examples and illustrations here: blog.alexewerlof.com/p/sli-compass

How do you know it's time to change jobs? What signals do you look for?

Incompetent leaders surround themselves with "yes people" and seek advice from external management consultancies. #leadership #ceo #cto #upperManagement #truthSeeking #growthMindset