Thursday, 19 November 2026

J-Fall 2026 - Writeup

So, I decided to just write some thoughts on some of the sessions I attended, as well as some bookmarks below.

Modernizing a 2M+ LOC Java Monolith with Agentic AI: What Worked, What Failed, and How We Recovered - Marek Dominiak
An interesting high-level talk (without code or demos). They used AI to provide a new implementation of an existing feature that was showing its age, without some of the disadvantages of the old implementation. It also showcases what can go wrong when trying to use AI, even with proper tests. They tried different AIs and they all provided very different solutions. They used Agentic AI in the end (see [6])
Why am I here? - Jeroen van der Wal (Keynote)
What if AI writes code for me. What is there for me? Because writing code is a thing every software designer has somewhere in their top three. They tried to have AI compete in the ICPC(See [7]) with pretty good results. If AI makes it easier to write software, we will write more software. See Jevons paradox8.
From Specification to Production: Engineering in the Age of AI - Venkat Subramanian
No matter how hard we try, we cannot get it right on the first try. Software development is an iterative process. Feedback is needed to iterate the software. Consequences are not always visible, until you try. AI makes it easier to iterate.
Hack Your Brain: Smarter Learning for Devs - Simone de Gijt
one of the best talks. It gave me some pointers on what's important to keep my mind concentrated and my body in a better condition. The slides are at [3].
Java over Markdown: Specs That Compile - Adam Bien

It turns out that the API Specs of Java and assorted technologies are very good for AI, because of the structure. Getting AI to generate an application based on APIs instead of implementations is likely to lead to good results.

Also specifying BCE (Boundary-Control-Entity (of Entity-Control_boundary ECB) Ivar Jacobson Object-Oriented Sotware Engineering: A Use Cae Driven Approach) seems to help a lot.

Let's get Physical - Iulia Feroli

An amazing and fun talk about Physical AI, which is a new buzzword for using AI in Robotics. An optimistic and engaging talk!

A colleague mentioned [1] as well.

Java 26 and 27 are boring, and that's a good thing - Lutske van der Meer, Johannes Bechberger
It's a boring talk, but has some highlights about things that I didn't know about Java 26 en 27. See [2]. But I'm waiting for Valhalla in Java 28 hopefully March 2027.
Java in 5 Microseconds: Inside a Crypto Trading Engine - Oleg Lobanov
Interesting talk. Basically, single-threaded application (it makes it easier to process orders in sequence which is a requirement in orderbook processing) that's optimised for speed can help a lot.
Conquest, time travel and explosions using Git - Alexander Chatzizacharias
An engaging and fun talk about how Git works, how you can use it to rescue Game Save Files, and how you can visualize git commits in a game. The demo was in fact a game.

I also took a look at the stands, and it appears JetBrains Kotlin has a new Mascot5 and I managed to pick up a fluffy toy of it.

Some sessions I would have liked to attend (hopefully I can watch them later on the youtube or some such).

  • Level Up Your Java Streams with Gatherers - Hinse ter Schuur
  • 6000 tests, 25 times a day, 365 days a year - All for quality - Sven Haster
  • Finding and Fixing Issues with Legacy Code using AI - Venkat Subramaniam
  • Apache Camel as integration framework - Ferrie Bank
  • Analytical Java with DuckDB - Geertjan Wielenga
  • Evolving Language Defaults in Kotlin and Other Languages - Michail Zarečenskij
  • 7 Ways You Can Build an Amazing Dev Career in the AI Era - Bruno Souza

References

[1] Robots that bring AI to life.
https://pollen-robotics.com
[2] Java 26 and 27 are boring and this is a good thing (talk)
https://github.com/parttimenerd/java26-27-are-boring
[3] 50m_HackYourBrain_DevoxxPoland_19062026.pptx
https://www.slideshare.net/slideshow/50m_hackyourbrain_devoxxpoland_19062026-pptx/288133705
[4] Code reading club
https://github.com/CodeReadingClubs/Resources
[5] JetBrains Blog - The Kotlin Mascot Returns – Say Hello to Kodee!
https://blog.jetbrains.com/kotlin/2023/04/the-kotlin-mascot-returns/
[6] Agentic AI
https://captainstory.gocaptain.ai
[7] International Collegiate Programming Contest (ICPC)
https://worldfinals.icpc.global/2026/about/
[8] Wikipedia - Jevons paradox
https://en.wikipedia.org/wiki/Jevons_paradox
The Big Blue Book - Domain Driven Design by Eric Evans
https://www.domainlanguage.com/ddd/blue-book/
Apache Camel – Enterprise Integration op zijn best
https://nljug.org/java-magazine/apache-camel-enterprise-integration-op-zijn-best/

No comments:

Post a Comment