04 Jun 2026 2 min read

Mission #2

A lot happened this week. An MLOps masterclass on killing brittle AI glue code, the Agentic AI Hackathon kicking off in SF tomorrow, and a hard look at why betting your architecture on short-lived AI SDKs is a trap. Let's get into it.

Agentic AI Hackathon - June 5, San Francisco

Most AI demos don't survive contact with production. This Friday's hackathon is built for pipelines that reason, backends that spin up themselves, agents that actually remember. Build a real agentic app using RocketRide, Butterbase, and xTrace in one day. Vaporware gets disqualified. Real integration wins prizes.

Claim your spot

Podcast: why fast AI still makes slow projects?

Coding agents are quick, no doubt. They also leave a mess. The real challenge now is what you do once the code part is easy. That's what RocketRide's founders dig into with Demetrios Brinkmann on the MLOps Community podcast, all from your IDE. Good one to put on.

Watch the Podcast

The AI layer is not your framework

There's a moment in every AI project where you sit back, look at the mess, and mutter "this framework sucks." But it really doesn't. You've just been duct-taping production onto LangChain and hoping nobody noticed. Joshua Phillips explains why your framework was never the problem, and what that messy middle layer is actually supposed to be.

Dive in