Vise Coding in Practice: Structured AI Development Across 5 Autonomy Levels | Blog This live session walked through five levels of AI coding autonomy in Java, but the real takeaway was simpler: sustainable AI development depends on small reviewable steps, explicit specs, and…

For anybody interested in the topic, but doesn't have the time to watch 2,5h of us talking, i tried to summarize a few key points in my blog post:
rabauer.dev/en/blog/vise...

Vise Coding in Practice: Structured AI Development Across 5 Autonomy Levels with Dr. David Faragó In this live coding session, I am joined by Dr. David Faragó to explore Vise Coding and the evolving landscape of AI coding agent autonomy.We break down and ...

You ever heard of Vise Coding?
I’m joined by Dr. David Faragó to put it into practice and explore the 5 levels of autonomous AI development.
All in Java, of course 😎
#AICoding #ViseCoding #Java #GitHubCopilotCLI
Join us on Thursday: youtube.com/live/7oVisn4...

Live Vibe Coding Battle: Build a Java App with GitHub Copilot | Blog Two predefined prompts, one shared Java app, and strict CI gates: the one-shot version started but failed when retrieving results, while the split-prompt workflow produced the more stable app and…

Hey people! I've decided to accompany my live coding session with blog posts to condense what i have learned in a session.
Check it out, if you can spare 5min ^^
Let's start with the session with @edelveis.dev
rabauer.dev/blog/vibe-co...

Neo4j + GraphRAG in Action: Live Refactor with Jennifer Reif In this live coding session I am joined by @jmhreif from @neo4j to refactor an AI powered file search application from PostgreSQL to Neo4j. The goal is to mo...

Going live today, 20:00 CEST 🚀

Jennifer Reif (@neo4j.com) and I are adding graph traversal on top of vector search and seeing what breaks (and what gets better). Expect a bit of chaos.

Join us 👇
youtube.com/live/KUcZL2y...
#GraphRAG #LangChain4J

Graph DB vs relational DB for AI apps… is it hype or actually useful?

We’re refactoring a real app live and testing it ourselves. Curious where graphs actually shine. Jump in and help us break it 🎭
youtube.com/live/KUcZL2y...
#Neo4j #GraphDatabases

I keep hearing about #GraphRAG vs classic #RAG …but what actually changes?

I’ll be figuring it out live with Jennifer Reif from @neo4j.com and asking all the naive questions. If you’ve got some, bring them here youtube.com/live/KUcZL2y...

Should be a fun experiment in disciplined AI collaboration.

Can AI help implement structured refactorings faster while keeping everything transparent and reproducible?

If you work with legacy Java, automated refactoring, or OpenRewrite, this might be interesting.

Nothing too outlandish.

The goal is not flashy AI code generation but deterministic refactoring. Every change will be defined as an OpenRewrite recipe that can be reviewed, versioned, and executed across repositories.

Plan for the session:
• Inspect a modernizable Java codebase
• Use @github.com Copilot to help write OpenRewrite recipes
• Apply those recipes to the project
• Analyze exactly how the code changes

AI Assisted Java Modernization: Building OpenRewrite Recipes with Tim te Beek In this live coding session I am joined by Tim te Beek from @Moderne-and-OpenRewrite to explore how AI can be used to modernize legacy Java applications in a...

Together with @timtebeek.github.io from @moderneai.bsky.social I’ll explore how AI can help modernize a Java codebase in a structured, reproducible way with @openrewrite.github.io

We’ll look at a real codebase and see how far we can push AI assisted refactoring.

youtube.com/live/LpGguXs...

Mit dabei sind auch @janekf.bsky.social, @gregorbiswanger.bsky.social, @johannesdienst.bsky.social, @richard.fichtner.dev und viele andere Speaker, die hier keinen Account haben 😎

Veranstaltungsbanner der Hochschule Weiden: ‚Dev-Stories mit Lerneffekt: Was schief ging und was half‘ am Freitag, 13.03.2026 um 10:45 in Weiden, Deutschland. In der Mitte ein Porträtfoto des Referenten Johannes Rabauer, der draußen an einem Laptop sitzt und lächelt.

An alle Oberpfälzer und die restliche deutsche Bevölkerung: 😅
Kommt doch gerne am Freitag zum #DCD in Weiden (an der OTH Weiden) in meine Session! Würde mich über jeden freuen und ich glaube für die Zuhörer wird's auch recht informativ und lustig 😎

www.dc-nordoberpfalz.de/DigitalCraft...

Looking forward to it! 🤓

So excited to do some vibe coding in a familiar environment, Java! :D
Regardless of the outcome: There will be something to learn from this. 🤓

If you're still vibe coding in 2026, you're leaving productivity on the table 😉

Watch the full session:
youtube.com/live/vopBYXp...

We also covered:
- Prompt injection risks
- Real guided coding projects
- Research workflows
- Live coding example

This is probably the most practical session I've been to on AI coding.

My favorite section was feedback loops:
- Automated tests
- Linters
- Benchmarks

Instead of guessing if AI code works, you prove it works.

AGENTS.md AGENTS.md is a simple, open format for guiding coding agents. Think of it as a README for agents.

The most practical part was AGENTS.md

A single file that tells AI agents:

-How the project works
-What matters
-Coding rules
-Constraints

It massively reduces hallucinations.

Key idea:
Vibe coding = hoping the AI gets it right
Guided coding = making the AI predictable

The difference is huge.

Guided coding uses:
- Clear plans
- Structured context
- Feedback loops
- Repeatable workflows

Many people are still doing vibe coding.

After a 2+ hour session with Kenny Pflug, I realized there's a better way: guided coding.

Less chaos.
More results.
Better AI collaboration.

Here’s the full session 👇
youtube.com/live/vopBYXp...

If you care about serious Java engineering, AI agents in real workflows, and maintainable code instead of quick hacks, this is for you.

AI and Java folks, join us live and bring your toughest questions.

Kenny will walk me through his Guided Coding framework:
• Plan first
• Constrain the AI
• Review aggressively
• Refactor deliberately

Using GitHub Copilot in Visual Studio Code with a real Java example.

Guided Coding instead of Vibe Coding in Java with Kenny Pflug In this live session we focus on a structured approach to AI assisted development: Guided Coding in Java with Kenny Pflug (@KennyPflug) Guided Coding was dev...

Live session: Guided Coding in Java with Kenny Pflug.
Not vibe coding. Not blind prompting.
We explore a structured way to collaborate with AI while keeping architectural control and code quality high.
youtube.com/live/vopBYXp...

To be honest: Never heard of it ^^

But after reading up on it, it does sound interesting, especially in the java environment. The tools i use usually have a single file, where they try to describe the structure of the codebase, but it seems like @buildwithbrokk.bsky.social does more here.

java.evolved Code Snippets | java.evolved A collection of modern Java code snippets. Every old Java pattern next to its clean, modern replacement — side by side.

java.evolved – Every old Java pattern next to its modern replacement, side by side.

javaevolved.github.io

Controlled AI Comparison in a Java Project: Copilot, Claude Code, Junie In this live session I run a controlled comparison of three AI coding agents inside a real Java project.The project is a complete chat application stack:- Va...

If you care about serious Java workflows, agent based development, and reproducible AI evaluation, this will be interesting.

Join live.
Give hints.
Challenge assumptions.
Or just watch what actually happens.

👉 youtube.com/live/XNtd3hJ...

#Java #SpringBoot #AIAgents #Copilot #Claude #IntelliJIDEA

We will measure:

• Time to working build
• Number of iterations
• What extra context the agent requests
• Whether tests pass
• Code quality and architectural discipline

If the model is identical, will the agents still behave differently?

The task is simple and strict.

Fix the failing tests.
Make the build green.
Make the application behave correctly.

That is it.

Three agents.
@github.com Copilot. @anthropic.com Claude Code. @jetbrains.com Junie.

Same project state.
Same prompt.
Same model: Claude Opus 4.6.

No architectural redesign. No shortcuts. No manual rescue unless clearly stated.

This is a full stack chat application with authentication, persistence, frontend integration, Docker, and failing integration tests.

Existing architecture. Existing tests. Real constraints.