Laravel News Interview and more

Laravel News Interview with Me about LLMs!

https://www.youtube.com/watch?v=YRG_eZJ0CnQ

One thing that came after the interview was a conversation with Eric and we found more then one way to help automate some of his more tedious daily tasks.

For example he gets numerous URLs sent to him daily that he has to verify are not spam. With just a little LLM prompting we can do the following.

Get the page from the url
See if the page is a valid technology article related to Laravel, PHP etc
If it is then mark the link as approved
if not mark the link as spam!

Of course it will take some prompt tweaking to get it right but in time as trust is built up he will no longer have to do this task.

NoteBookLM makes a Podcast!

This is an amazing tool and everyone who wants to see how LLMs can really help them do things like "study a collection of documents", "create FAQs" from a collection of documents, or make a podcast! You can see here I release a few chapters and some PHP LLM news (coming soon) https://readorlisten.com/episodes/why-llms-chapter-2-of-php-and-llms-the-book

The Best RAG Technique Yet? Anthropic’s Contextual Retrieval Explained!

This has shown up in a few areas and it is starting to take off. Augmenting your RAG system with chunk context!

https://www.youtube.com/watch?v=tmiBae2goJM

GraphRAG: The Marriage of Knowledge Graphs and RAG: Emil Eifrem

Unlock the connections in your data with a well known pattern of Graph Databases and then using RAG to let the LLM help you query that data. https://www.youtube.com/watch?v=knDDGYHnnSI

AR Glasses

Not only does Zuckerberg release Llama 3.2 multimodal so now we can upload images to this open-source model but he showed off AR glasses

https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/

Interesting interview here https://www.youtube.com/watch?v=8dVba_xm4MQ

GraphRAG: The Marriage of Knowledge Graphs and RAG: Emil EifremUnlock the connections in your data with a well known pattern of Graph Databases and then using RAG to let the LLM help you query that data.

https://www.youtube.com/watch?v=knDDGYHnnSI