Ryan Castillo returns to talk about pivoting Knowatoa from an SEO play into an AI-visibility tool — and why the analytics everyone wants to track are becoming impossible to trust in an AI-search world. The back half turns into a deep dive on Ryan's AI coding workflow: skipping the Claude chat interface entirely in favor of firing off cloud agents from Cursor, letting them run unsupervised, and reviewing PRs cold instead of living in the conversation. Plus: why "throw away the code, not the plan" beats hand-holding an agent through every step, and why the AI labs have zero incentive to ever let you leave the chat.

Links:

- Knowatoa

- Ryan Castillo's X/Twitter

- Ryan Castillo's LinkedIn

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

Ernesto Tagwerker (FastRuby.io / OmbuLabs) returns to talk through how AI is reshaping both sides of his business — speeding up Rails upgrades while simultaneously eating into revenue, and opening a new lane in AI strategy consulting for non-technical businesses. The group digs into the messiness of AI-generated tests, mutation testing, and how to know when your tests are actually testing anything.

  01:00 FastRuby's Open Source Claude Skill for Rails Upgrades
  03:30 How AI Has Eroded (and Grown) FastRuby's Revenue
  07:00 Engineering as Marketing — Why They Open Sourced the Playbook
  10:00 The Tooling Isn't the Problem — Engineering Culture Is
  12:30 Using Claude to Write Contract Tests and Improve Coverage
  15:30 The Problem with AI-Generated Tests
  17:00 Mutation Testing Explained
  19:00 Classicist vs. Mockist — Where Claude Gets It Wrong
  23:00 Playwright for Acceptance Testing in the Development Loop
  26:00 Who Watches the Watcher?
  28:00 Staff Augmentation and AI Strategy at OmbuLabs
  31:00 Working with Non-Technical Businesses in Philadelphia
  33:30 Engagement Sizes — AI Discovery vs. Rails Upgrade Projects
  35:30 CRM Summaries and One-on-One Prep with AI
  38:30 Observability and Guardrails — How Do You Know It's Working?
  43:00 LLM as Judge
  45:00 Event Scraping — AI Makes Scrapers More Maintainable
  50:00 Ryan's Universal Scraper — Using AI to Heal, Not Replace

LINKS

- Ernesto's LinkedIn

- Ernesto's Bluesky

- FastRuby.io

- OmbuLabs

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Send us Fan Mail

Colleen Schnettler and Joe Masilotti (Permission Not Required podcast) join Ryan and Brendan to catch up on what they're each building. Joe walks through Ruby Native — his platform for getting Rails apps into the App Store without touching Xcode or writing native code — and how AI has changed his development process. Colleen shares what AI automation consulting actually looks like in practice: more process mapping and integrations work than pure AI, and why developer credibility matters in a crowded market. The group also compares notes on how they're each using AI beyond coding — for business context, client management, and personal workflows.

  00:00 Introductions
  01:00 Ruby Native — Rails to the App Store Without Xcode
  05:30 What Changes When You're the Only One Reading the Code
  08:00 AI Automation Consulting in a Crowded Market
  13:00 Beyond Content and Sales — Where AI Actually Fits
  16:00 Cowork, Skills Repos, and Dropping OpenClaw
  18:00 Joe's Three-Agent Coding Workflow
  22:00 Projects You'd Never Have Tackled Before AI
  24:30 Ruby Native vs Hotwire Native Explained
  28:00 How the Preview Workflow Works
  33:30 Leveraging Existing Networks for New Businesses
  41:00 When Automation Doesn't Need AI
  44:00 The Command Center and Second Brain Thread
  49:00 Joe's Business Context in Claude — Powerful and Dystopian
  51:30 Colleen's Personal Agents (School Emails, LinkedIn Pipeline)
  54:30 Where to Store Context — Markdown, SQLite, and Avoiding Beads
  56:00 Joe at RubyConf in July

LINKS

- Permission Not Required Podcast

- Colleen's LinkedIn

- Colleen's Email: [email protected]

- Joe's Website

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

Ryan and Brendan talk through the growing pains of AI-assisted development — including fatigue, increased confidence in bigger code changes, and what happens when you stop paying for expensive APM tools and just build your own. Ryan walks through how he used Claude to identify memory pressure, fix N+1 queries, and cut his Heroku costs significantly in just a few days. They also dig into the real tradeoffs: maintenance burden, over-optimization, and why you still need to understand what Claude is telling you before you run anything on a live database.

LINKS

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

Scott Werner (Sublayer) returns to discuss how rapidly AI and agentic coding have is changing.  The hosts and Scott dig into why niche tools are increasingly getting replicated by developers and hobbyists with "good enough" custom solutions — while complex enterprise platforms may actually have more staying power. Scott shares what he's seeing at the Artificial Ruby meetups. They also cover Stripe's "forward deployed" AI role, the reality of AI consulting, growing trust in AI-written code, and orchestration tools like Fabro for defining repeatable AI workflows.

01:03 Early Agentic Coding Days
04:49 SaaS Is Dead Debate
10:28 Enterprise vs Homegrown Systems
12:52 Forward Deployed Builders
15:20 Architects Still Matter
19:33 Personal Software vs Production
23:22 Stripe Forward Deployed AI Role
26:16 AI Adoption and Consulting Window
29:42 Games and What Software Becomes
33:20 Burnout and Pairing Lessons
36:06 Learning to Trust AI Output
38:57 Requirements Over Code Style
43:02 Multi Project AI Workflow
45:36 Software Factory Orchestration
54:40 Process Thinking and QA
56:41 Skills and Non Determinism

LINKS

- Sublayer Website

- Scott's Blog (Works on My Machine)

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

The hosts explore whether “data is the new moat” in SaaS and argues that data alone—especially customer data—is rarely defensible as extraction and migration get easier. Instead, moats increasingly come from clear product vision, differentiated approaches, integrated systems, and how well apps enable AI and humans to interact with structured data. They contrast niche SaaS with broad platforms like HubSpot and Salesforce, discuss pricing model changes driven by AI costs (moving beyond per-seat to metered/credits), and predict reduced reliance on traditional UIs as chat and CLIs become efficient interfaces for AI-to-backend communication.

00:00 Data as a Moat
01:52 Niche vs Platform SaaS
02:45 Integrations and UX Moats
05:31 AI Changes Pricing Models
06:48 Geocoding Data Example
08:35 When Not to Use AI
10:43 Why CLIs Are Rising
14:25 Future Without Interfaces
18:43 Data Product vs Customer Data
21:28 Opinionated SaaS Wins
23:19 AI Change And Strategy
23:48 Defensibility And Platform Risk
25:19 Building A Command Center
27:30 Top Tools In The Stack
27:52 Do You Still Need UIs
30:36 Second Brain For Priorities
33:37 Frameworks And Memory Systems
36:02 SaaS Moats In An AI World
39:23 AI Beyond Tech Businesses
41:35 Business Takeaways And Wrap

LINKS

• Ryan's Website
• Brendan's X/Twitter
• Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

The hosts once again discuss the recurring theme of AI’s impact on development workflows.  They discuss various topics, including: 

• Increased PR volume from AI-assisted coding and whether traditional pull request reviews are still necessary. 
• Should AI be used earlier for pre-PR review and stronger CI guardrails for performance and security. 
• What is the role of automated testing, and are they even more important as code generation accelerates?

01:05 PR Reviews Under AI Load
02:38 What PR Review Is For
05:02 Human Review What Remains
08:36 AI Pre Review And PR Purpose
11:01 Guardrails CI Performance Security
12:39 TDD Tests As Behavior Spec
15:53 AI Versus Automation In Pipeline
19:17 Testing Evolution Unit Vs System
20:32 Feature Specs And Functional Review
21:35 Bugs vs Defects Focus
23:37 Smoke Tests for Upgrades
24:44 Controller Tests Debate
27:40 Test Pyramid and Frontend Gap
28:37 Refactoring for Unit Tests
30:57 AI Coding Raises Review Stakes
34:00 Claude MD Friction
35:54 Wrap Up and Testing Habit

LINKS

• Justin Searls Podcast
• Working Effectively with Legacy Code Book by Michael Feathers
• Professional Rails Testing: Tools and Principles Book by Jason Swett
• Ryan's Website
• Brendan's X/Twitter
• Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

In this episode, the hosts discuss building a “universal scraper” for event calendars and a new AI-assisted workflow using a long-running orchestrator.md thread.

00:12 Why Change AI Workflow
01:29 Orchestrator File Setup
04:36 How Memory Persists
07:33 Platform Constraints WSL
08:45 Git Worktrees and Isolation
12:07 Scraper Build Approach
18:26 Incremental TDD Development
20:17 Making It Truly Universal
22:39 Universal Scraper Explained
26:03 Wrap Up and Next Steps

LINKS

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

The hosts discuss how AI-assisted coding has massively increased their output, creating a new bottleneck: cognitive load, PR review, testing, and getting work shipped. Ryan discusses splitting time between finishing near-complete work, advancing a few large architectural initiatives, and handling ongoing support.  Brendan talks about his productivity system using Claude to manage his time using local markdown files to help stay on top of things. They also cover risks of AI automation, permission scoping, subagent/approval bugs, and conclude with a renewed optimism that AI raises abstraction, increases competition, and pushes SaaS to adapt via execution, moats, and AI-ready APIs.

00:27 AI Code Volume
02:17 Cognitive Load Shift
04:52 PR Buckets Strategy
07:22 Overwhelmed Yet Excited
10:36 Workflow Tools Check
11:51 Claude Productivity Setup
14:02 Second Brain In Practice
17:21 Simple Local Files
19:30 Automation Limits
22:13 Scheduled Script Automation
23:00 Claude Cowork vs Command Center
23:39 API Permissions and Safety
26:12 Claude Code Approval Friction
29:25 Subagents and Parallel Work
30:44 When Agents Pretend Work
34:34 AI Shifts SaaS Thinking

LINKS

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail

Today we discuss lessons from our previous episode with John Nunemaker, focusing on the Conductor tool and how multi-workspace, multi-agent workflows can speed up coding and bug fixing. Brendan shares his experimenting with parallelizing six Rollbar error fixes at once by generating prompts and running them in separate Conductor workspaces. They compare this with a one-agent-at-a-time workflow and debate whether closed-loop setups (ports, databases, dependencies) could make parallel work trustworthy and reduce context-switching costs. They also cover using full Honeybadger/XML reports for faster debugging, using Claude Code from a phone to create PRs, challenges syncing session history, and broader AI product strategy like APIs/MCP, and RAG.

00:00 Catch Up And Recap
00:43 Conductor Workflow Overview
02:40 Parallel Rollbar Fixes
04:35 Manual Testing And Setup Hurdles
06:33 Debugging With Full Reports
08:46 Ports Docker And Dependencies
11:50 Parallelism Versus Focus
14:57 Closed Loop Trust And Context
18:24 Merge Conflicts And Acceptance Gaps
22:16 Review Bottleneck And Output Surge
25:26 Fixing Gallery Uploads
26:39 Multi Select Challenges
28:27 Branching Without Conductor
31:28 Claude Mobile Workflow
32:34 Session Sync Friction
34:44 AI Brain For SaaS
38:01 APIs And MCP Table Stakes
41:49 Internal AI Assistants
43:46 Access And Safety Concerns
46:31 Second Brain Revival
49:03 RAG Tooling Experiments
50:04 Wrap Up And Listener Feedback

LINKS

- Ryan's Website

- Brendan's X/Twitter

- Brendan's Bluesky

Questions or comments, email us at [email protected]

Send us Fan Mail