Two weeks ago, Model Context Protocol shipped its 2026-07-28 specification revision.

The biggest updates in this release are: no more stateful initialize handshake, Mcp-Session-Id headers have been retired, and every request …

Have you heard the classic story about a county fair where people were guessing the weight of an ox? Nobody was exactly right, but somehow the average of all the guesses was remarkably accurate! Francis Galton wrote an …

I created a skill to save users 40-60% on token costs when migrating AI inference workloads from a serverless platform (like Cloud Run or Gemini Enterprise Agent Platform) to GKE. So what is it? What isn’t it? Why a …

Introduction to Distributed RL Sandboxing on GKE

Reinforcement Learning (RL) is the cornerstone of modern AI training. Rather than train a model to produce an expected output, we verify if it has achieved a particular outcome. This is …

The Am Dash And Discerning Human Writing from AI

Was This Post Written by Gemini?^*

The other day I was listening to one of my favorite podcasts talking about AI’s writing style. And of course, they were talking about the Em dash. I’ll let …

Fine-Tuning Mixtral-8x7B: A Deep Dive into a Multi-Node Slurm and FSDP Workflow

Training or fine-tuning large language models like Mixtral-8x7B is a significant challenge due to their large size. Sometimes, to make these …

Fine-tuning Gemma 3 on an A4 Slurm Cluster

Note: This post has been adapted into a formal tutorial on the Google Cloud documentation site.

Overview

This post demonstrates how to fine-tune the Gemma 3 language model on a multi-node, …

Fine-tuning Gemma 3 on an A3 Mega Slurm Cluster

Note: This post has been adapted into a formal tutorial on the Google Cloud documentation site.

Overview

This post demonstrates how to fine-tune the Gemma 3 language model on a multi-node, …