Everyone uses AI. Few use it well. Between prompts written with no real criteria, agents handed tasks a simple script would do better, and teams “adopting AI” without changing a single process, the gap between hype and real productivity keeps widening.
At Codemotion Milan 2026, a group of speakers takes on that gap with no easy answers: no polished demos, just what actually happens when AI meets real code, real teams, real decisions. Here are seven sessions worth blocking off in your calendar.
When the Agent Is the Wrong Answer
Everything Is Slop, Let’s Slow the F*ck Down starts from a blunt observation: the coding agents that felt like magic on a side project become a complexity nightmare in production. You’ll learn why agents amplify bad habits at industrial speed, what “slop” really means as it floods open source PRs, and how to recover discipline and control without giving up AI’s real advantages. Practical knowledge you can apply to your repo on Monday morning, before it becomes a problem. The speaker is Mario Zechner, one of our headline names.
On the architecture side, When NOT to Use an Agent tears apart the reflex of “need automation? throw an agent at it.” You’ll learn to read the real trade-offs between deterministic workflows, service architectures, and agentic orchestration: blast radius, cost, observability, attack surface. Not a feature shootout, but a mental framework you’ll reach for every time someone in a meeting suggests “let’s just put an agent on top.” Presented by Jigyasa Grover, also among our top international speakers.
Closing the trio is the mythbusting Why Coding Agents Need You More Than Ever. No “developers are obsolete,” no “it’s just autocomplete”: the argument is that an LLM has no internal sense of correctness and no memory across sessions, and filling that gap takes exactly the design and engineering skills many assumed were becoming obsolete. You’ll learn why capable agents make software engineering more necessary, not less, a grounded reassurance about the value of your career, not wishful thinking. Delivered by Alfonso Fuggetta.
Adopting AI Without the Illusions
Scaling AI Adoption is the story of trying to move 300 engineers past the “messing around with ChatGPT” phase toward repeatable, AI-native development. You’ll learn what AI-Native Engineering and Spec-Driven Development look like at scale, but more importantly what doesn’t work: resistance, false starts, over-reliance on tooling, and the gap between perceived and actual productivity. A blueprint, scars included, valuable if AI adoption on your team is still a patchwork of individual initiatives. Shared by Alfonso Graziano.
Kill the Standup opens with an extreme case: nineteen thousand lines added solo to Node.js’s Virtual File System in three months. You’ll learn how autonomy and taste, combined with an AI that removes the throughput bottleneck, can let one developer outship an entire team, and what happens when one of those two ingredients is missing (spoiler: a CVE). A sharp take on Conway’s Law and the hidden cost of Scrum in the agentic era, useful whether you lead a team or just want to see where your own role is heading. Written and delivered by Matteo Collina, one of the best-known names in the Node.js ecosystem.
How to Interview Engineers in the AI Era takes the conversation somewhere few look: the technical interview itself. You’ll learn a practical framework, built from hundreds of interviews redesigned from scratch, for assessing how someone reasons, collaborates with AI, and applies judgment when the model’s output is incomplete or wrong. Useful whether you’re hiring or prepping for an interview: the metric you’ll be judged on has already changed. Explained by Dennis Nerush.
The Tools Rewriting the Rules
Green Prompting flips the usual framing of prompt engineering: not just “write better prompts to get better answers,” but write prompts that cut AI’s energy footprint. You’ll learn concrete, practical instructions anyone can apply to reduce token consumption and cost without sacrificing answer quality, a skill that will soon show up in enterprise project requirements, not just ESG slide decks. Guided by Valeria Salis.
The Revolution of Embedded AI in Browsers brings AI somewhere nobody expects it: inside the browser itself. You’ll learn how Chrome is integrating native models into its runtime, what consensus is forming across browsers on the future of Web AI, and what you can already build today with the Chrome AI APIs, ahead of what the speaker calls Web 4.0: the agentic web. A real competitive edge for anyone building web products who wants to get there before everyone else. Presented by Erick Wendel, Node.js core committer, Google Developer Expert, and Microsoft MVP, with over 100 talks delivered in more than 20 countries.
Seven talks, one common thread: AI isn’t a magic button. It’s an architectural, organizational, ethical, even environmental choice, and it deserves to be treated as one. Anyone who walks out of these rooms at Codemotion Milan 2026 won’t have heard more hype. They’ll have a method.
