We’ve all been there: “This is the third time I’ve been asked this — I should really write it down.” Then comes the sigh, because writing it down means reconstructing all the context, and for what? No one reads documentation until they need it — and by then, it’s usually out of date. Even when it isn’t, no one wants to wade through a long doc for one answer. They’ll just ask you the question directly anyway.

Based on Andrej Karpathy’s “LLM Wiki”, this talk is about ending that cycle with a pattern for LLM-maintained documentation — one where the LLM keeps the wiki updated and answers questions directly from it, citing and filing each answer back in for next time. Internal docs don’t go stale because no one cares; they go stale because maintenance cost outpaces the value of doing it by hand. Once an LLM can do that maintenance and answer the question itself for near-zero cost, “always up to date” stops being aspirational.

In the fast-paced world of software engineering, processes often shape success or sow frustration. Over the years, recurring anti-patterns have emerged well-meaning practices that unintentionally create inefficiencies, stifle innovation, or harm team dynamics. These pitfalls, though rooted in good intentions, can derail projects and undermine trust when left unchecked.

Through real-world examples and two decades of industry experience, this presentation unpacks common process anti-patterns across the software lifecycle. With a focus on fostering understanding, not blame, we’ll examine how flawed workflows contribute to challenges and how reversing them can empower teams to build clarity, creativity, collaboration, and trust.

Rails apps usually reach for external AI services when they need “intelligence,” but that approach can bring latency, cost, and unpredictability along for the ride. This talk shows a different path: running a Japanese sentiment model directly inside a Rails app, using ONNX Runtime, background jobs, and real-time updates to keep the user experience fast and responsive.

Using a small app called Sentiment Omikuji as the case study, the talk walks through how the model was trained in Python, exported to ONNX, and loaded into Ruby for local inference. It also covers how Japanese text is tokenized, how the sentiment output drives a fortune generator, and why the app uses Solid Queue to keep model work off the request path. Along the way, the talk explores the tradeoffs of local inference versus API-based LLMs, and why classic machine learning can still be the right tool for focused problems.

Attendees will leave with a practical architecture they can adapt in their own Rails applications, plus a clearer picture of when local AI makes sense, how to keep it predictable, and what it takes to make it feel good in a real app.

Pull requests are where quality is won or lost, and too many teams still treat reviews with a rushed “LGTM”. In this talk, you’ll learn a practical, repeatable approach to high-quality PR reviews that balances speed with risk management - covering how to triage changes quickly, review in layers, and write comments that lead to better outcomes without friction. We’ll also show how QA and developers bring complementary lenses to the same review, turning acceptance criteria, scenarios, and observability into shared responsibility instead of production surprises. You’ll leave with a lightweight framework and concrete habits for authors and reviewers that make reviews faster, kinder, and more effective.

The networks we use every day depend on centralized infrastructure that’s vulnerable to censorship, surveillance, and single points of failure. Whether it’s disasters, deliberate interference, or plain old downtime, entire populations can, in an instant, be cut off from digital communication—sometimes when they need it most. The heart of the problem is hierarchy—the hierarchical structure of these networks makes them easy to exploit or to simply shut down.

Reticulum is one solution to this problem. It’s a cryptography-based networking stack that flattens the hierarchy entirely. It was built for creating resilient, ad hoc, and private networks using almost any communication hardware you can imagine—from LoRa, WiFi, and Ethernet to old school packet radio and serial ports. Whether you’re dealing with natural disasters or just want communications that can’t be easily disrupted, Reticulum is a powerful tool for building independent, interconnectable networks that work without central points of control.

This talk will break down what Reticulum is and dig into how the protocol works—from its clever addressing tricks to its baked-in encryption. We’ll explore the built-in utilities that ship with Reticulum; try out apps like MeshChat, NomadNet, and Sideband for communication; and even show some code and learn how to build apps of our own.

By the end, you’ll understand how Reticulum and the applications that run on it work and how to set it all up for yourself. Plus, you’ll have the knowledge to grow the network and its capabilities by writing with applications of your own. But most importantly, you’ll walk away understanding how to build communication networks that can survive disasters, resist interference, and keep working when traditional infrastructure fails.

Backed-up job queues, crashing worker processes, memory bloat? If you’ve had to support production background job workloads, you’ve felt it. Give me five minutes, and I’ll show you the five simple steps we’ve used repeatedly for resilient job queues—queues that are cost-efficient, fault-tolerant, and run smoothly without manual intervention.

When Active Record works, it’s great.  Polymorphic associations in Active Record promise elegant flexibility: one interface, many possible models. When they work, they feel almost magical. But when your application grows and real‑world complexity sets in, that magic can quickly turn into confusion, broken queries, and performance headaches. We’ll explore where polymorphic associations shine, where they silently introduce technical debt, and why even experienced Rails developers can end up wrestling with unscoped queries, tricky joins, and hard‑to‑optimize relationships. We’ll walk through real examples of tangled polymorphic logic—and more importantly, show practical strategies to refactor, untangle, or even avoid these pitfalls entirely. By the end, you’ll know when to reach for polymorphism, when to run away from it, and what to do when your complex query simply refuses to cooperate. -

Let’s talk about what makes Ruby on Rails able to Full-Stack Flex! This talk was given at CodeMash 2026 and I’ll talk about bringing Ruby and Rails back to CodeMash!

Hi all, given that there’s a level 3 snow emergency in Franklin County, we’re going to cancel tonight’s CRB and resume in Feb. Hope everyone’s staying safe and warm!

Presented by Tech Community Coalition

Join the Columbus tech community for our annual holiday celebration! This is a great opportunity to connect with fellow developers, designers, and tech professionals from across the city.

🎟️ Details and Your Free Tickets on Eventbrite

• When: December 11th, 2025
• Location: 421 West State Street
• Cost: Free

CRB will meet next in January