Cheezy World

Delivering software with discipline and joy

Strides New Enhanced Workflow

Over the past two weeks, a lot of enhancements have been made to the Stride Workflow. I wanted to take the time to detail them here as I think they are a good example of how one can direct an AI agent to perform specific tasks in a repeatable manner.

What is Stride?

For those new here: Stride is a kanban-based task management platform built specifically for Human and AI agent collaboration. It is very different than the typical “let the agents run through a lot of tasks and we’ll check at the end” mode of working with AI. Instead, Stride gives the human (team) multiple places where they can insert themselves into the workflow. Enough of that for now as I want this post to focus on the workflow enhancements.


Teaching AI Agents to Look Before They Leap

The Stride Claude Code plugin just shipped version 1.1, and the headline feature is subagent orchestration — the ability for an AI agent to dispatch specialized sub-agents at key points in the task lifecycle. Instead of one agent doing everything sequentially, the primary agent now coordinates a team: an explorer that reads the codebase before coding starts, and a reviewer that checks the work before tests run.

This post covers what changed and how it works.


What Impact Will AI Have On Outsourcing?

I was in a conversation with a potential new client last Friday, and they told me that they still had some of their development outsourced to an offshore company. After the call I started to think about the outsourcing model. How will it be impacted by AI? It led me to a conclusion that might be uncomfortable for some.

Why did companies outsource?

In my experience, there have been two primary reasons for companies to outsource development:


Building the Learning Management System - First Coding Session

This post is a continuation of the implementation of the Learning Management System application demonstrating my concept for the new team structure and process in the age of AI-driven development.

In the first post of the implementation series, we created a requirements document with the whole team and Claude Code. In the second post, the whole team, using Claude Code, created a backlog of tasks from that requirements document and uploaded it to Stride. The entire process and team structure is outlined in the post titled “The New Team.” If you haven’t read the initial post, I recommend you do so before reading this one and follow along with the implementation via the links at the bottom of that post.


Claude Remote Control

Anthropic is in the process of releasing a feature for Claude Code that I think is amazing. It is Claude Remote Control. This feature allows you to control your Claude Code session remotely from any iOS or Android device.

I can see this feature being used in many different scenarios. For example, you could use it to check on work while you are in a meeting, to kick off a task while you’re at lunch with friends, or to simply sit on the couch and oversee some work Claude is doing. I used it to get a few tasks completed while having coffee with my son. It was extremely easy to use and I can tell you I will be using it a lot. I personally tested it on both my iPhone and iPad and it works great.