0

Join my free GitHub Actions Security Workshop on July 14th, 2026!

Homepage
Courses Membership The Guild
Content
Newsletter Blog YouTube Podcasts Discord Server GitHub Examples Sponsor
Consulting
Private Training Workshops
About
About Me Link Tree
Log In
← Back to all posts

My Favorite AI tools, local models, MCP hacking

by Bret Fisher
Oct 06, 2025

A new Agentic DevOps episode on my latest favorite AI news and tools, plus, join me next week for a special live stream with Cast AI on live pod migrations between K8s nodes.


🎧 Agentic DevOps Podcast Release

My Favorite AI Terminal, Prompt Injection, and More

(Audio episode only)

Agentic DevOps | My Favorite AI Terminal, Prompt Injection, an...

In this episode, I walk though some of my favorite new AI tools and content.Resources and Topics mentioned in this episode:My AI Skeptic ...

agenticdevops.fm

Here are some topics from the show

  • My AI Skeptic Friends are All Nuts, blog post from fly.io by Thomas Ptacek. An interesting set of arguments for why developers should be using AI code generation.
  • From Toil to Triumph: Harnessing Agentic AI to Streamline Infrastructure as Code, KubeCon London 2025 video by Jody Varney. This was the only talk at the conference that addressed what it truly takes to build a reliable PR review agent. (The talk is 8 months old now, so it's likely gotten easier since.)
  • The Dark Side of Just Hooking Up AI Agents to GitHub, blog post by Simon Maple from AI Native Dev. GREAT example of how easy it is to accidentally send malicious prompts to someone else's AI (prompt injection.)
  • Warp 2.0 My new daily terminal. I love the design of having a built-in agent and a shell prompt that dynamically switches between cli commands and agent prompting. Warp 2.0 walkthrough
  • Container Use project - A good idea for how to ensure your local agents always edit/write changes in a container to reduce the blast radius of project changes. This also allows you to have multiple agents running on the same codebase without interference.
  • Devstral release in May - Devstral and Qwen-3-Coder seem to be the best two models for local code generation. I can run them on my Mac, and they are designed to fit on a single desktop GPU or Mac. Here is the latest Devstral release and the Qwen-3-Coder. For Mac, I use them from LM Studio, which supports the MLX format that is faster on Mac than GGUF. (I'd use Docker Model Runner, but they don't support the MLX format yet)
  • SWE Bench for finding the best model + tool combos for AI code generation.

Live Stream this Thursday: K8s Pod live migrations between nodes

πŸ™Œ Special Event on Thursday, October 2nd: Learn about Kubernetes Pod Live Migrations (between nodes) with me and Cast.AI, and maybe win our prizes, including AirPods Pro 3!

Container Live Migration: Moving Workloads Without Downtime - ...

Learn live container migration in Kubernetes with Cast AI and Bret Fisher. Hands-on demo, real-world cases, and expert insights

cast.ai

 

My GitHub Security Hardening workshop is free next week
Sign up for my live hands-on workshop that's on Tuesday. I've been working on it for months, then I spoke about it at the Accelerate Chicago conference last month, and I finally get to bring it to you wonderful people. I found this content so useful for my own repos and orgs, that I created an open-source tool to complement the training and help you harden your repositories and find where you ...
AI-Native Kubernetes: What Does it Mean?
Join my stream Thursday 6/4/26 with guests from OpenChoreo, a new CNCF Kubernetes "AI abstraction" with built-in agents and all the fixin’s for managing K8s with agents. KubeCon has been talking about AI in every keynote since the launch of ChatGPT, but it wasn't what you think. Before that moment in consumer AI, the language of AI was MLOps-specific and rarely highlighted at KubeCon. Soon aft...
The AI cost flip-flop
More "devil in the details" of AI engineering is starting to surface. Subscriptions are getting more expensive or offering less, model APIs have frequent outages, costs are skyrocketing once engineering teams fully adopt it, AI has limited effects on productivity due to it exposing downstream bottlenecks (PR reviews) that take humans to solve, and even "firing AI and hiring devs" was reported l...

Cloud Native DevOps

CNDO is a weekly-ish email from Bret Fisher on content he's creating. Agentic DevOps, automation, containers, Docker, Kubernetes, GitOps, GitHub Actions, and everything cloud native.
Homepage
© 2026 bretfisher.com
Courses Membership The Guild
Newsletter Blog YouTube Podcasts Discord Server GitHub Examples Sponsor
Private Training Workshops
About Me Link Tree

Join Our Free Trial

Get started today before this once in a lifetime opportunity expires.