One of my favorite album of all time is 20 years old today ❤️

In a near future (or maybe it's already happening) we will waste a huge amount of time debugging AI-generated code or even reading emails written by LLMs.

Modern GPUs can already generate all that crap in a blink, so they will call it a 'dramatic productivity increase'.

'hallucinations are an *inherent feature* of AI large language models'

In other words: artificial intelligence is not a replacement for human one, so don't trust it blindly.

futurism.com/the-byte/ceo-g…

Pushing Java to the Limits: Processing a Billion Rows in under 2 Seconds by Roy van Rijn

youtube.com/watch?v=9-S_nZ…

Roy van Rijn #jvmperformance #1BRC

Generative AI Conversations using LangChain4j ChatMemory | @SivaLabs bit.ly/3uIONwk
#langchain4j #Java

Repeat after me:

'I do not have to integrate AI into everything.'

Many tech companies are going all-in into the AI hype, partially putting aside other well established and profitable businesses and products for no valid reasons. A relevant fraction of them will repent this decision.

Rule engines: the thing that we called artificial intelligence before it was cool!

Why do men go into bars to meet women instead of entering to Zara? There are many more women and they are already looking for stuff they don't need.

Processing CloudEvents with Drools by Toshiya Kobayashi blog.kie.org/2024/05/proces…

How many times have you been perplexed when users report issues, but instead of pasting the code they used, they add screenshots?

With OpenAI's GPT 4o and Quarkus LangChain4j it's now so easy to get the code from those screenshot.

Don't believe me? See:

github.com/geoand/quarkus…

Some of my (non-technical) friends started using ChatGPT as a normal search engine and I'm having quite a hard time explaining them what's the difference and why in many cases this is not a good idea. Are you experiencing similar situations?

I recorded a video on how to run #AI LLM models (such as InstructLab) with Podman AI Lab & then call the LLM from a Java/Quarkus app. All running on my local machine and #OpenSource , of course 😁 youtube.com/watch?v=Zogg2c…

Podman QuarkusIO #PodmanDesktop Red Hat Developer

Quarkus Insights #166: Lies, Damn Lies...and Techempower(?) x.com/i/broadcasts/1…

A demonstration of ML + Symbolic Reasoning using Quarkus, Drools & LangChain4j by Mario Fusco 🇪🇺🇺🇦

github.com/mariofusco/qua…

The pain of maintaining an open source project 😢

80% of tasks that we associate to general intelligence consist in some sort of pattern recognition.

LLMs do pure pattern recognition and they are becoming very good at it.

Nevertheless pattern recognition is only part of the story and to achieve real AGI we will need much more.

Would you ever buy a car (or anything else) from the guy who had the brilliant idea of renaming that into this?

Seen in Bairro Alto, Lisbon.
💯% agreed!

Start of a long weekend in Lisbon to celebrate (with a bit of delay) my 50th birthday with the friends of a life.