Dear Phoenix Friends,

We’ve entered a new era—one that doesn’t wait for instructions.

Phoenix demonstrated a major step forward in AI-powered software development: a shift from chat-based AI to closed-loop AI. Unlike traditional chat interfaces, closed-loop systems don’t wait for you to type a prompt. They don’t require hand-holding or supervision. They’re always on, always working, always learning.

Phoenix generated a complete Rails test suite—start to finish, with 98.62% test coverage and zero failing tests—and it did it without a single line of human-authored test code. When it got stuck, it retried automatically. It switched between six different LLMs to get the job done. It never tapped a shoulder or waited for a Slack message. A task that took a team of two humans over two months was completed by Phoenix in just a few hours.

But this isn’t just about speed. It’s about care.

Phoenix doesn’t just write tests. It nurtures your application. It identifies the highest-risk parts of your codebase. It flags what needs the most attention. It generates specs for untested code, finds bugs, and offers test-driven fixes. Over time, Phoenix creates a feedback loop where your application gets stronger, more stable, and more joyful to work on—week after week, commit after commit.

You can visit phoenix.love to learn more, sign up for this newsletter, or schedule a demo to see it in action.

(This is Issue 16. Missed one? Find them all here.)

On the Technical Side

Update:

As we scale Phoenix for larger codebases, our focus has been strategic. We’re working on ways to identify high-value entry points—places in the codebase where Phoenix can deliver maximum ROI on Day 1. That means blending the new with the classic.

Yes, we’re pulling out flog, flay, and even heckle, combining them with our AI tooling to inform how and where Phoenix begins its work. The result is a smarter, faster onboarding process—and more meaningful outcomes for customers.

Challenge:

Larger applications introduce more complexity, but also more opportunity. The trick is helping Phoenix choose its battles wisely. By pairing smart heuristics with old-school insight, we’re teaching Phoenix to prioritize with intention.

How You Can Help:

Have experience taming large Rails monoliths? Favorite tools for analyzing legacy code? Drop us a line—we’re always looking to sharpen our approach.

On the Business Side

Update:

We’ve onboarded two more companies—and the queue is growing. We’re expanding our onboarding capabilities so more teams can experience the joy of working with Phoenix, faster.

We’re also thrilled to share that Joe has launched a podcast with Ruby legend Valentino Stoll. It’s coming soon to your favorite podcast platforms, and it’s full of real talk about Ruby, AI, software craft, and the future we’re all building together.

Challenge:

Demand is rising. Now our mission is to scale without losing what makes Phoenix special: the deep care, responsiveness, and love that have defined our journey so far.

How You Can Help:

Know someone who’s been burned by brittle test suites or struggling with Rails complexity? Point them to phoenix.love. We’d love to show them how Phoenix can help their app rise from the ashes—with more coverage, fewer bugs, and a lot more joy.

❤️ The Phoenix Team