Raphael De Lio

Redis 8 just made KEYS and SCAN faster and safer

Operations that used to take 12–14 seconds on large datasets can now complete in just a few milliseconds — making previously discouraged commands viable for in production.

Read the full article at: redis.io/blog/faster-...

If the third era is about autonomous execution, then the real question might be: are our evaluation systems, testing practices, and engineering standards ready for it?

Interesting article by the founder of Cursor:

x.com/mntruell/sta...

O mais interessante é que literalmente nenhum argumento histórico ou técnico que essa pessoa fez tem qualquer embasamento. Até a analogia do primata foi ruim. É mesmo de impressionar.

Can’t wait to be back at @springio.net !! The conference that made me love conferences!

@anthropic.com says Claude helped NASA plan a 400m drive for Perseverance on Mars.

For AI engineers, the lesson is: it was not a bigger model. It was context.

Clear constraints. Domain knowledge. Structured outputs. Simulation. Human review.

Reliability comes from system design.

Anthropic just announced they're giving 6-month free of Claude Code 20X for open source contributors!

If you maintain a popular project or contribute across open source, you should apply!

claude.com/contact-sale...

But do you implement an alternative in prd? Or is it simply not necessary?

I got the 8bitdo adapter too. Just a shame it won't work with the NSO NGC Controller

Workshops don't need to run simultaneously and they're very lightweight. But what I'm trying to anticipate is that I want to make an online version of this so that learners won't even need Docker to go through the labs. I wonder if DinD would fall short in a production environment?

How does it work for production? Does it spin up other pods in k8s?

I'm currently using it in a workshop hub I'm building to make it easier for learners to discover, deploy, and go through hands on labs by themselves. All they need would be Docker. Not even an IDE. It's working fine for now.

When does DinD (Docker in docker) start falling short and when is it just enough?

What is your experience with DinD (Docker in Docker)? Would you recommend it? What use case did you try to build with it?

Docker-in-docker powered by SpringBoot. How cool is that?!

A complete hands on lab with no IDE required, no environment setup, no git pull, and no friction.

Just open it and start building. #javabubble

Love that KLM now let us use their planes’ wifi with no limitations for free! Now I can vibe code at 30,000 ft - currently working on my social media tracker agent built with Spring AI and Redis

For future reference: When using Gemini models, we're required to explicitly enable structure outputs in the ChatClient. I was also using Spring AI Vertex AI integration instead of Google GenAI. Thanks @tzolov.bsky.social and @ddobrin.bsky.social!

This was something I discovered while trying to build a reliable multi agent with Spring AI fully within GCP. With OpenAI it works flawlessly. With Gemini, constant structured output issues... Has anyone else seen this? #java #javabubble

Am I the only finding Gemini to be less consistent to adhere to structured output than OpenAI?

#SpringAI Tool Search Tool - Build smarter AI apps that discover tools when they need them

📖 Blog: spring.io/blog/2025/12...
🔗 GitHub: github.com/spring-ai-community/spring-ai-tool-search-tool

✅ On-demand tool discovery
✅ Multiple LLMs
✅ 34-64% token savings

@spring-ai.bsky.social

Uma das tradições de natal é comer KFC no Japão 😄

“No thoughts go unpublished.” Literally 😆

Kick-Off Belgian Kotlin User Group, Thu, Dec 11, 2025, 6:30 PM | Meetup Hi Kotlin enthusiasts! I am so hyped to be hosting our first BEKUG Meetup at Kunlabora. This will be an important first time event where we'll welcome you all as members

In 2 weeks we’re kicking off the Belgium @kotlinlang.org User Group with @sch3lp.bsky.social!

The first session since 2019 😍

We’re starting off with a talk, and afterwards we’ll have an open space.

There’ll will be lots of Kotlin and Kodee goodies 😁 See you there!

www.meetup.com/belgian-kotl...

What happens when you get Kotlin, Spring AI, Redis, and data from the Apollo 11 together? A whole new way of exploring one the most important milestones humanity has ever accomplished!

Kotlin conferences just hit different! And it was a pleasure to be part of Kotlin Dev Day! @kotlinlang.org

JSON is the go-to format for LLM tool responses, but alternatives like TOON, XML, CSV, YAML can potentially improve token efficiency and performance.
Find how to configure #SpringAI tool responses formats and optimize for your use case:
spring.io/blog/2025/11...
@spring-ai.bsky.social

I also have Codemotion Madrid and KotlinConf closing soon in my system #javabubble

I don't even know anymore if LLMs are dropping the quality of anything produced by humans in every field (not only software development) or if everything has always been as mediocre as it is today. We're shipping faster and in more quantity, sure, but clearly at the expense of quality.

This week is all about Kotlin as I get ready to speak at Kotlin Dev Day!

Can't wait for a day full of Kotlin talks led by a few of the best engineers I have had the pleasure of meeting.

If you still haven't got a ticket, get one with 20% discount with the code: kotlinfriends@kdd25 #java #kotlin

All my fellow @kotlinlang.org server devs, there is still time to fill the annual Ktor survey!

If you're using Ktor then please consider filling the survey, and help us continue making it better.

I went to the @mlopscommunity.bsky.social conference as an attendee today, but when I learned anyone could get on stage and deliver a lightning talk, I didn’t hesitate! 7 minutes to let the audience know how semantic caching can optimize agentic ai costs!

Why You Need Semantic Routing in Your LangGraph Toolkit: A Beginner’s Guide Not every question needs a 175-billion-parameter LLM. Semantic routing helps decide which model, tool, or path should handle each query

Semantic routing improves predictability and efficiency in AI systems. RedisVL classifies user queries without LLM calls, reducing token usage and latency. Combined with LangGraph, it creates a deterministic flow for routing, execution, and decision logic. Read how here: medium.com/@bhavana0405...

Two more weeks to submit your talks for @kotlinconf.com! In case you’ve been procrastinating like me.. This weekend might be the perfect time to write yours too!

Come and share any awesome work you’ve done with @kotlinlang.org! Good luck 👍

sessionize.com/kotlinconf-2...