Speed isn't Enough - How to Code with Velocity Using AI (Ruby Montevideo)

This Ruby Montevideo meetup features talks on the intersection of Ruby, Rails, and modern development practices. In this session, I will be presenting an updated version of my talk "Speed isn’t Enough," specifically tailored to the rapid developments we've seen in the first months of 2026.

Speed vs. Velocity

While AI can generate code at incredible speeds, the real challenge for developers is maintaining velocity—which is speed with a clear direction. Without a structured process, using AI can feel like navigating a first-person shooter without instructions. We will delve into how to use the OODA loop (Observe, Orient, Decide, Act) mental model to stay in control of the AI-augmented development process.

What's New in 2026?

The landscape has changed significantly since the original version of this talk. This update includes:

• The Age of Agents: How OpenAI's multi-agent systems and Anthropic's "Computer Use" are shifting AI from reactive chat interfaces to proactive digital collaborators that can navigate entire development environments.
• Advanced Reasoning & Deep Thinking: A look at the Gemini 3.1 Pro and Deep Think modes, and how they help solve complex architectural problems that simpler models struggle with.
• Autonomous Security and Patching: Discussing how tools like Claude Code Security are autonomously identifying and fixing vulnerabilities, and what this means for the future of code reviews.
• Modernizing with AI: Strategies for using AI to tackle technical debt and modernize legacy codebases while maintaining the high standards expected in the Ruby community.

The goal is to provide developers with a practical framework to integrate these power tools into their Ruby and Rails workflows while increasing both speed and quality.