Vibecoding Meetup — Agentic Coding

This time, it’s all about Agentic Coding!

Whether you are a pure vibecoder building entire apps with natural language or a seasoned dev architecting autonomous agent loops, we want to hear your story.

The session is designed to be accessible for newcomers, while still providing depth and practical value for people already working with AI tools.

🕒 Date: May 19th, 18:00
📍 Location: TBD

Agenda: TBA

Confirmed Speakers:

Edouard Maleix - How AI-First Dev Teams Build Collective Intelligence
Edouard has been watching teams accumulate agent rules, notes, and feedback controls that never quite add up to a real system. Agents repeat mistakes across sessions, guidance grows without being validated, and what gets built is instructions, not reusable knowledge.

His talk is about closing that gap: giving agents their own identity, signed commits, and reasoning linked to every change. When something breaks, the agent captures it, links it to the fix, compares it with lessons from other agents and humans, and feeds what survives back into future work.

The goal is making mistakes compound into collective intelligence instead of disappearing into chat history.

***

Andrew Demczuk - Claws Across the Internet
Andrew is an OpenClaw contributor and AI battle arena builder who has spent his time figuring out what happens when you stop thinking about single agents and start connecting agents with agents.

His talk is about AWA: Agents with Agents. When one agent can securely spawn and coordinate subagents across the internet, the force multiplier changes the game entirely. Andrew has stress-tested this in some of the more demanding environments imaginable: competitive AI battle arenas like CodeClash and BitGN, where agents don't just assist, they compete.

The talk covers what agentic coordination actually looks like in practice, what secure subagent spawning requires, and what builders can take from the competitive AI space into their own workflows.

***

David Wischnewski - Vibe Coding vs. Agentic Engineering
David isn't a developer. He's a builder who uses AI to turn ideas into working tools, workflows, and products. That puts him in an interesting position to compare two distinct ways of working with AI: vibe coding, which is fast and chaotic, and agentic engineering, which starts to become reliable.

In this talk, he shares what he's learned building with AI as a non-engineer: where coding agents genuinely help, where they break, and which habits separate a fun demo from something you can actually trust.

***

Karim El Kaddioui - AI Agents in Knowledge Work
Karim is the FP&A Team Lead at Meister, and he's increasingly adding terminal-based AI agents into new and existing workflows. Using Codex, Claude Code, and Claude Cowork, he runs data analysis, builds decks, and manages his task tracker without touching most of the traditional tooling.

This talk is a practitioner's view of agentic coding outside of software development: what it actually looks like when a finance professional runs agents in the terminal, where it works well, and what it takes to get there.

***

Daniel Siegl - The Worktree Multiverse: Where AI Agents Build Software
Daniel Siegl is talking about infrastructure: specifically, the Git primitives that agentic development needs.
Worktrees and Git notes have been around for years and mostly ignored. Agentic software engineering makes them essential. When autonomous agents work in parallel, managing state, history, and collaboration across sessions requires more than a single checkout and a commit message.

This talk revisits these underused Git concepts and explains why they're suddenly at the center of orchestrating agents that build software.

***

Want to speak at our next event? The topic of our next meetup is Vibecoding in Marketing. How do marketers of all types use vibecoding in their roles?
Apply ➡️ https://forms.gle/z3GwEheRycJGMfZ6A