The modules run from how an AI model produces an answer to running AI inside a live operation. If AI is new to you, start with AI fundamentals and read in order. If you have a specific decision to make, go straight to the module that covers it. For a quick definition of a single term, the glossary covers the ones that come up most, each linked to the chapter that explains it.
AI fundamentals
What's actually happening when AI answers: where the words come from, why some of them are made up, how much it can hold in its head, and why the way you ask changes what you get.
- What an AI model actually is, and how LLMs work →
- Why AI hallucinates (makes things up) →
- What a context window is, and what happens when it fills up →
- How to write a good AI prompt, and why prompts matter →
- How AI learns: the 4 stages of training a model →
- Where AI gets its information, and what it actually knows →
- AI vs. machine learning, and where LLMs fit →
- How ChatGPT and other AI assistants work, step by step →
The AI tools landscape
The AI you buy, the model underneath it, and the platform it runs on are three different things. Who makes what, and how open and closed models divide the market.
- Is ChatGPT an LLM? Models, products, and platforms →
- Which AI model is best? GPT, Claude, Gemini and the rest →
- Open-source vs. closed AI models: what is actually open →
- Multimodal AI: image, audio, and video →
- Choosing AI: the 4 questions behind every working system →
What AI can and can't do at work
What AI does well at work and where it fails, how it reaches company systems and takes action, and how to test it before trusting it.
- What AI can do well in business operations →
- The limitations of AI at work, and why its mistakes get past review →
- How AI uses tools and memory →
- How AI connects to company systems, and what MCP is →
- What an AI agent is, and what changes when AI takes action →
- Is AI good at coding? What the evidence shows →
- RAG vs. fine-tuning vs. prompting, and what 'train it on our data' means →
- AI wrappers: what gets built around the model →
- Testing AI before trusting it: how to run an eval on your own work →
AI risk & trust
Where your data goes, who owns what AI produces, what quietly goes wrong, and the controls that keep people in charge.
- The 4 risks of using AI in business →
- Is ChatGPT safe for company data? Where your data goes →
- AI and copyright: who owns the output, and what is safe to publish →
- Human in the loop and guardrails: how much checking AI needs →
Putting AI to work
How AI lands inside a real operation: spotting where it fits, build versus buy, what a good project looks like, and how you know it's working.
- How to find AI use cases in your business →
- Where AI fits in a business: which tasks it can help with →
- How much AI costs, and how token pricing works →
- Build vs. buy AI: how to decide →
- Why most AI projects fail, and what makes a good one →
- How to measure AI results and ROI →
- Running AI in production: cost, security, governance and drift →
Comet Lab writes Horizon and adds chapters as AI changes. To suggest a correction or a plainer way to say something, write to hello@cometlab.ai.