michaelisvy

Migrated my @SpringAICentral project to 1.0.0-RC1. Here are the commits (breaking changes in the POM file, MCP support and SimpleVectorStore) github.com/michaelisvy/de…

@starbuxman it has been a while my friend :). But you can disable the video if you just need the audio

guuuuys did u know Spotify did video ?????? HOW COME NOBODY TOLD ME THIS

just upgraded my @SpringAICentral project to 1.0.0 M7. Mostly POM and package changes. Here is the changelog in case anyone needs it github.com/michaelisvy/de… Excited that it's going GA next month!

@BriceDutheil plugins.jetbrains.com/plugin/20540-w…

@michaelisvy Oh no !

wow this is cool, IntelliJ is finally catching up! Windsurf is coming to IntelliJ (and other Jetbrains products) with in-editor changes, so we don't need to copy/paste from the chat window anymore windsurf.com/jetbrains_tuto…

@therealdanvega yes, I agree price shouldn't be the main differentiator. It will be much more if you use it full time. Also those are subsidised prices, will be much more in 1-2 years when those companies will need to make money.

I feel like when we talk about Cursor vs Windsurf pricing is a big part of the conversation. Cursor for $20 a month where Windsurf is $15 a month but you can quickly eat up credits and need to move to $60 a month. Is this on par with everything you're seeing as well?

really cool article about Copilot versus Cursor. I realise there are a few Cursor features that I didn't know builder.io/blog/cursor-vs…

Moving my small demo app to Spring AI 1.0.0 M5 was smooth. Only had to change the call to VectorStore. Next step is to try out the new audio support. github.com/michaelisvy/de… #springai

@piduruviswa @cursor_ai @bolt yes this field is moving very fast. We'll be doing a SingaDEV meetup on that on Jan 15th if you'd like to come, building some apps from scratch using Cursor and Copilot

@michaelisvy @cursor_ai I have tried @cursor_ai for iOS app prototyping and it was good. I have been tinkering with @bolt and it looks promising for full stack development

been playing a lot on building Spring Boot applications with @cursor_ai Experience is flabbergasting : code refactoring support, diffs are possibility to cancel changes are awesome. VSCode is so much ahead with coding assistants. Wondering if IntelliJ is going to catch up

@gcuisinier @juliendubois hello Gildas ! merci pour l'invit. Je te fais signe si jamais j'ai l'occasion de passer au Luxembourg

@juliendubois @michaelisvy Vous voulez pas revenir le faire au Luxembourg aussi ? Ca aurait été cool de vous revoir vous deux mais ca fait loin ;-)

Join us in Paris for a LangChain4j + Spring AI event! meetup.com/nightclazz/eve… I'll be delivering the LangChain4j part and my old friend @michaelisvy will do Spring AI. Thank you so much @ZenikaIT for hosting us!!

@sergialmar Happy birthday Sergi! Welcome to the club

I'm turning 40 today 🎉 Looking forward to the next decade and what’s coming next!

@rey_antoine @odedia @juliendubois @agoncal @glaforge great news!

The Spring Petclinic ecosystem has been enhanced with two new Spring AI and LangChain4j chatbot versions: github.com/spring-petclin… github.com/spring-petclin… Thanks to the main contributor @odedia and for the hands-on on LangChain4j @juliendubois @agoncal @glaforge #DevFestNantes

@emmanuelbernard I would tend to say that prompts are like SQL. I've been adding them to the code

In RAG apps, are you in the camp 1. all my prompts in one file because a dedicated person works on them 2. Put my prompt next to the code, it’s so intertwined, I would hate the back and forth Explain why in response for extra ♥️

@therealdanvega it's true, but the TokenTextSplitter is still a bit limited. I hope there will be more options soon.

Spring AI makes RAG effortless! 🚀 Add your own documents / data to a PgVector Database and then add context to your AI request with just one line 🔥 Instant improvement to your AI's knowledge. Try it now at start.spring.io! #SpringAI #RAG #Java

@therealdanvega hi Dan, by the way another topic that might be interesting to cover (but maybe a bit more advanced?): Ollama is a development tool (doesn't support multi-threading, doesn't support models more than 13B etc). How to switch in production into something such as LlamaCPP?

@therealdanvega Some topics I struggled à but to find info about: - chatmemory with conversation id (so each user has her own conversation) - doing RAG for excel and pdf files with Apache tika integration

🔥Planning out some more videos on Generative #AI in #Java with Spring AI 🤖Did you know it is super easy to get started with and use features like streaming & RAG with just a few lines of code? What topics would you like to see me cover? github.com/danvega/spring…

I just saw that Spring AI 1.0.0 M3 has been released. Upgraded my demo application to it, everything worked out of the box github.com/michaelisvy/de…

@habuma super clear, thanks! I see we should use conversation id. I'll try it out

@michaelisvy How’s this? github.com/habuma/spring-…

I've said this before in my Spring AI sessions and even used the analogy in my book. But in case you missed it: The movie "50 First Dates" and how Henry reminds Lucy everyday about what has transpired describes perfectly how chat/conversational memory works in Generative AI.

@habuma thanks Craig. I agree, but didn't find any good example on how to make sure that each user has their own session. For instance, I can expose a Rest endpoint and use the InMemoryChatMemory. But how to make sure that each of my users has their own session?

Spring AI makes that easy, though with 3 different chat memory advisors.