2025: A Final Reflection

As we reach the end of 2025, the question is no longer what Artificial Intelligence is capable of doing, but how we choose to live alongside it.

AI has ceased to be merely a technical promise or a set of well-placed buzzwords in corporate presentations. In 2025, it began to generate concrete, measurable, and often irreversible consequences.

This was the year in which AI came to occupy a role closer to that of a team member than that of simple software. Not only executing tasks, but helping to structure ideas, explore alternatives, and support decisions in increasingly complex contexts.

Companies and professionals who understood this shift did not just produce more, they produced better.

Because when AI stops being just a “tool” and starts acting as a thinking partner, the gain is not only in speed, but in the quality of decisions.

Newsletter Highlights

I’d like to use this space to briefly look back on our newsletter. 🎉

Throughout 2025, we published 50 articles, exploring topics such as RAG, AI agents, LLM training and adaptation, multimodal models, and other subjects in a technical and applied way for those who work or conduct research in the field.

Today, the newsletter has nearly 1,800 subscribers and a collection of 34 notebooks with practical examples, all designed for direct execution in Google Colab, making experimentation, learning, and reproducing the concepts discussed in the articles easier.

Below are some highlights from the newsletter in 2025:

• MCP: Anthropic’s Open Protocol That Is Connecting AIs to the Real World

• The Best Path to Master Artificial Intelligence – A Complete Guide (Part 1)

• LoRA: How to Adapt LLMs Efficiently and Quickly

• Everything You Need to Know About LangChain (Before Starting an AI Project) – Part 1

• 9 Language Models with the Best Cost–Benefit for Running Locally

• How to Run Hugging Face Models Directly in VS Code with Copilot Chat

• Nanochat: A ChatGPT Clone for Just $100

• Multimodal RAG in Practice with Open-Source Models

• Jupyter AI: Turning the Notebook into an Intelligent Lab

• How to Build Your Own Local RAG with Wikipedia Content and LEANN

Thanks and holiday wishes!

Thank you to everyone who followed, shared, and helped expand these discussions throughout the year!

Feel free to connect with me on LinkedIn and Instagram as well, where I continue sharing analyses, reflections, and updates on AI.

Let’s keep going together in 2026! 🚀

Happy holidays and a great new cycle ahead!!