Rails Business

No guest this week — just Brendan and Ryan working through several things they've been seeing lately. Topics covered include:

  - Anthropic's skills: why both of us started by writing one giant skill instead of small composable ones, and why that's the same lesson object-oriented programming taught us years ago.

  - Where skills end and agents begin — plus a surprise from Brendan's code review setup, where the sub-agents it spun up weren't using the skill at all.

  - Ryan's spam war at Locable: 2,000 fake accounts before breakfast, 60,000 businesses stuck waiting on human review, and whether an LLM should get a vote on what's legitimate.

  - Why the real blocker isn't the skill — it's the APIs, MCPs, and evaluation systems underneath it that nobody talks about.

  - And the PR backlog problem: AI gets you 90% there in an hour, the last 10% is as hard as it ever was, and Intercom and Shopify keep their pull requests under 200 lines for a reason.
LINKS
Anthropic Developers Skill Usage Video
Intercom 2x Shipping Blog Post
Ryan's Website
Brendan's X/Twitter
Brendan's Bluesky
--Sponsor--
This episode is brought to you by Flagrant.  Your partners for delivering thoughtful products & applications from any stage in your project.  Book an hour with Flagrant at beflagrant.com/call
Questions or comments, email us at railsbusinesspod@gmail.com
Send us Fan Mail

Jesse Hanley joined us from southern Japan, where he runs Bento — an email marketing and transactional email platform — as a company of one. Topics covered include:
Why the bootstrapped dream now means picking big, hairy, competitive markets instead of cozy niches AI can replicate overnight.
Automating entire roles with agentic "employees" — Cursor Cloud Agents and Grok Bot wired into MCPs for support, compliance, infra, finance and dev.
Why most jobs are really just cron jobs, and how far one person can go if they build the tools deep enough.
Plus Heroku scale-to-zero, moving Postgres off to PlanetScale, and why he's eating his words on agents replacing labor.
LINKS
- Bento
- Tatami
- Ryan's Website
- Brendan's X/Twitter
- Brendan's Bluesky
--Sponsor--
This episode is brought to you by Flagrant.  Your partners for delivering thoughtful products & applications from any stage in your project.
Questions or comments, email us at railsbusinesspod@gmail.com
Send us Fan Mail

Jim Remsik — founder of Flagrant, Madison Ruby and the traveling XO Ruby conferences — joins us for a wide-ranging conversation.  The discussion starts with the ups and downs of running a design-and-dev agency in the AI era.  Topics covered include:
- Why design has started outselling development at Flagrant.
- Why accessibility and UX are the parts people forget, and how Flagrant's hiring process mirrors real work to surface the thing that matters most: communication.
-  Plus the story of XO Ruby — small, one-day, city-to-city conferences built to make community accessible again.
LINKS
- Flagrant
- XO Ruby
- Ryan's Website
- Brendan's X/Twitter
- Brendan's Bluesky
Questions or comments, email us at railsbusinesspod@gmail.com
Send us Fan Mail

Nate Berkopec (Speedshop) returns to dig into what AI actually changes about Rails performance work — and it's not what the hype says. The through-line: stop conflating understanding code with verifying it, because the real leverage is building objective, tool-callable checks an agent can loop against — N+1s, characterization tests, jemalloc — instead of shipping more unreviewed slop. Nate makes the case that the future looks like Intercom's "software factory" (tiny, automatically-verified PRs, most merged with no human review) and asks why the rest of us are still hand-building 4,000-line monstrosities. Plus: why "intelligence" is the wrong word for what these models do, and a first look at Thread Pilot, his new automatic thread-pool tuner for Sidekiq and Puma.
LINKS
- Speedshop
- Nate's X/Twitter
- Threadpilot
- The Complete Guide to Rails Performance Book
- Sidekiq in Practice Book
- On Rails Episode with Brian Scanlan from Intercom
- Ryan's Website
- Brendan's X/Twitter
- Brendan's Bluesky
Questions or comments, email us at railsbusinesspod@gmail.com
Send us Fan Mail

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 railsbusinesspod@gmail.com
Send us Fan Mail