Before anyone shared a single line of code, Dana Mansolillo explained why she needed to build something. That moment, unscripted, specific, and quietly brave, set the tone for everything that followed at February’s Code & Share, the monthly virtual meetup hosted by Code Connector. What unfolded wasn’t a demo night. It was four people explaining the wound before they explained the tool.
The Idea That Hasn’t Shipped Yet
Dana Mansolillo opened by naming something most builders keep private: the daily friction of being neurodiverse in a world of productivity tools built for someone else’s brain. Six apps (ToDo, Toggl, Reminders, and more), each solving part of the problem, none solving all of it. What she wanted to build was a single tool that understood executive function: routines, due dates, stakeholder tagging, a persistent top-five task view, time blocking, context filtering, and calendar integration.
The project doesn’t exist yet. That’s not a caveat. It’s the point.
Dana walked through each existing app’s strengths and the gaps she kept falling through, with the precision of someone who has been quietly cataloging frustration for a long time. The audience responded with something specific: not polite encouragement, but recognition. A few shared their own experiences. Others praised her sharpness of vision. The room understood that it was watching someone tackle the hardest part of building: identifying the real problem before reaching for a solution.
Code & Share makes space for the idea that hasn’t shipped. Dana’s presentation was a reminder that the “why” behind a project (the lived need, the felt absence) is often the most important thing a builder can bring to a room. And in this room, it was enough.
The Tool That Built Itself
John Hall arrived with a different kind of vulnerability: the intellectual kind. He had needed an AI agent that could branch conversations in ways commercial tools couldn’t. When nothing on the market fit, he didn’t adapt his workflow. He built infrastructure.
He called it DogFood, an agent that builds agents. What he presented was the idea for DogFood v2, built using DogFood v1. The architecture carried its own logic: sub-agents for research, building, and analysis; GitHub integration; a CLI and chat UI; local or cloud deployment. The “Human in the Loop” framing wasn’t offered as a limitation. It was offered as an intention. Cost control, yes. But more precisely: goal alignment. The human stays in the loop because the human knows what the work is actually for.
The recursion of it was the statement. John didn’t just dogfood his tool; he used it to become a more capable version of the builder who made it. For anyone in the Memphis tech community navigating the gap between what AI agents promise and what they actually deliver, his session made something concrete. When the tools don’t fit, building your own is a legitimate path forward.
Research-Grade Work, From a Phone
Kevin Kawchak came to Code & Share with a GitHub repository built entirely from a mobile device using AI. Focused on physical AI cancer trial research, generated using Claude and ChatGPT, complete with Mermaid diagrams, changelogs, READMEs, and a clean directory structure. The kind of output that signals rigor before anyone reads a line.
What drove it was curiosity and method. Kevin had studied how high-profile AI-created repositories were structured, then reverse-engineered those best practices into his own workflow. The Q&A that followed, on token efficiency and LLM selection, treated prompt engineering the way it deserves to be. As an iterative craft. Not a one-shot command.
The absence of a workstation wasn’t a constraint Kevin worked around. It was evidence of something the session made visible. That serious AI-assisted research is accessible to anyone willing to study how the best builders approach it. For Memphis developers still calibrating their relationship with AI tooling, Kevin’s presentation lowered the assumed barrier without lowering the standard.
Two Projects, One Origin
JC Smiley closed the afternoon with two updates, and the room shifted for the second one.
The first was the Memphis Tech Scene website: an updated event calendar, new newsletter sections, and an expanded blog section. Ecosystem infrastructure. The kind of work that holds a community together without announcing itself.
The second was Chapters. JC described it plainly: a healing-through-storytelling app for cardiac trauma survivors, built after his own heart attack. The personal origin wasn’t framed as a backstory. It was framed as the reason the design decisions matter, the reason getting it right is non-negotiable, the reason the app exists at all.
What the room heard was a builder bringing his own recovery into the work. That kind of transparency doesn’t just humanize a project. It raises the permission level for everyone else in the room to bring their whole reason into what they’re building. Dana had done it first. JC did it last. The afternoon ended where it began: with someone explaining the wound before explaining the tool.
What Code & Share keeps modeling, month after month, is something that doesn’t happen by accident. Technical people, in a room together, telling the truth about why they’re building. Not as a warm-up to the real presentation, but as the presentation itself. The work matters. But in this community, the work only fully lands when the room understands what drove someone to start. That’s not a format. That’s a culture. And it’s one of the quieter, more durable things Memphis tech has built.
Memphis Tech Is Compounding.
Every week, new events surface. New partnerships form. New lessons are shared publicly.
Stay connected to what’s actually happening, not just what’s announced.
95+ Issues Published
600+ Events Amplified
250K+ Digital Impressions