Ever wondered why the prompt size is limited when prompting an AI model, or why it seems to forget what it was told earlier in the interactions? Context windows provide working memory to AI.

RAG helps AI move beyond its training limits by dynamically retrieving relevant information snippets, allowing language models to answer specific questions with precision and context.

Weight, parameter and tensor are used when talking about LLM model sizes, are they all the same thing?

Small AI models are cost-effective, fast, and ideal for business tasks. Cheaper to train and run, they work locally, offline, or in IoT devices, ensuring efficiency and compliance. With less energy use and fine-tuning options, they prove that smaller, task-focused AI can deliver big results.

It is fun to download Ollama and run an AI model locally on your laptop or desktop computer, even without an Internet connection. It teaches us to think about model selection even in cloud AI.

Logic of thought prompting is a technique used to force the AI out of speedy responses and in to making a more considered answer with significantly more thought and detail.

Academic and business growth of AI and recent news on Amazon giving academics more compute.

Open weight models vs open source models, they are often confused when people discuss AI models.

Guardrails facilitate responsible AI, guiding users to good results from AI and putting a barrier up to prevent harm. Guardrails are the safety barrier used by AI systems to ensure the system remain on target for their intended use and do not cause harm.

By learning prompt techniques, you gain the ability to tailor AI outputs to your specific needs, making interactions with AI more productive, whether you're brainstorming ideas, generating content, analysing data, or seeking insights.

AI Agents, Agentic AI, Agentive AI, AI Assistant, CoPilots are all terms we see around business automation. All relate to using AI to help accomplish tasks but with varying amounts of human intervention and drive.

Having a model in your head when working with generative AI and LLM platforms is a great help. This posts paints an overview of AI models and introduces common terms every new comer to AI platforms encounters.

0-Shot, 1-Shot, 2-Shot, 3-Shot, 4-Shot, 5-Shot are all a way of prompt engineering, using examples to shape the response from an LLM.

When downloading or selecting models, you will see them named instruct, chat, chat-instruct models, what are they and what are the differences?