Here’s my latest post on AI with 60+ worthwhile links found since my last post on it in June. It comes in three sections: the first section is about AI technology, the middle section is on AI in the news and the oped pages, and at the end there’s a good but odd bunch worth a look. Enjoy!

Technology – LLMs:

• This is a good repo on tuning Llama2.
• This talks about the LLM llama3.
• Another good repo:  Phi 3CookBook here. More on phi3 here. Also here.
• This repo deals with neo4J and LLMs.
• Not all language models are small. For example, Reader-LM is a small language model.
• Here’s something on LLM pre training and post training.
• This here repo is all about building an LLM from scratch. Really good material.
• Last, here’s a guide to build your own RAG app: a step by step guide to setup LLMs locally using Ollama, python and chromadb. Nice.

Technology – Ollama / Flowise:

• If you need to know how to run ollama in a docker environment, this is  your complete guide.
• For people wanting to  get started with Flowise.
• Advice on how to  integrate  google for flowise.
• A piece on  Flowise and Langchain.

Technology – RAG:

• A simple repo on how to put together RAG and Ollama.
• On moving from from RAG to agent systems.
• Here’s a recipe for RAG – how cloud services enable generative ai outcomes across industries.
• For haters of RAG: in defence of RAG in the era of long-context language models.
• Here’s how to harness rag for text, tables and images
• If you want master RAG and select an embedding model.

Technology – watsonx.ai:

• First off, a watsonx.ai quick start guide.
• Two good piece on setting up  watsonx .ai to work with python, here and here. Relatedly, this is also very good.
• On setting up of watsonx.ai in the cloud.
• Go here for the watsonx.ai Prompt Lab.
• Here’s info on watsonx.ai’s foundation models.

AI in the news / oped sections:

• Looks like  Mcdonalds had to end their AI drive thru test. Ah well.
• Speaking of AI fail:  Detroit facial recognition software had led to false arrests.
• More AI fail:  Jenna Ortega quit Twitter due to fake AI images being generated of her. Sad.
• Here’s a piece in the New York Times on  AI chatbots and language learning models. The Times also wrote that the AI boom has an unlikely early winner :wonky consultants.
• There’s been a lot of focus of how AI and new data centers are going to cause climate change issues due to a rise in carbon emissions. Relatedly, here’s why AI and clean energy need each other.
• On how AI is being used to catch tax fraud.
• On improvements of  ChatGPT in math and science.
• A fun article by Kash Hill who let generative ai make all the decisions in her life.
• A piece on how the Quebec government is using AI to track trees, pools, etc, here.
• Looks like the “ignore all previous instructions” loophole in  chatgpt gpt 4o is being closed. Good.
• How AI blew up nanowrimo.
• Oops! Google AI is telling people to put glue in pizza. Oh dear.
• Here’s all of Dave Karpf’s critiques of IT and the industry. Smart stuff.
• Sad stuff: AI took their jobs. Now their job is to make AI sound more human.
• On the AI bubble,  here.
• On AI agents.
• On AI slop.

Last but not least:

• This is a really good repo / course on generative AI for beginners. (and check out these other repositories here, too).
• A good intro to embedding, here.
• A tool I really like: autogen studio.
• How to create an Azure AI resource.
• If you need a ChatGPT cheat sheet, go here.
• A good list of  prompt engineering tools
• Here’s what you need if you want to  understand deep learning
• On  AIops vs Mlops.
• This here on the importance of responsible AI.
• A step by step guide to using generative AI in your business, here.
• A helpful piece on dealing with AI hype, here.