About this guide
An independent guide to help anyone who works in an office understand what generative AI is, what it can do, and how to use ChatGPT, Gemini, Grok and Claude well at work. A GenAI.mil edition (CAC) for uniformed and civilian members of the U.S. military exists at sean.theinformed.org/gen-ai/; this open edition is for everyone else.
Not affiliated with OpenAI, Google, xAI or Anthropic, and nothing here overrides your organization's rules. The views are the authors' own.
Who wrote it
Sean Egan is a U.S. Navy information warfare officer and a research scientist. He holds a PhD from the University of Alaska Fairbanks, where he modeled and remotely sensed volcanic-ash hazards to aviation. He writes code and builds tools as well, but most of his time goes into training machine-learning models for weather on his own hardware: short-range radar nowcasting, satellite super-resolution and true-color imagery. He wrote the four explainer books linked from this site at sean.theinformed.org.
Derek Conklin is a U.S. Navy Aerographer's Mate First Class with the Strike Group Oceanography Team in Norfolk, after tours with the Naval Oceanography Mine Warfare Center and the Fleet Weather Center Norfolk aviation detachment. He does much of the coding behind these sites: the websites themselves, the tools and the scripts. He teaches meteorology, oceanography and information warfare, works on unmanned systems and machine learning for forecasting, and builds learning platforms including Weather Woobie at derek.theinformed.org.
Between them, one of us mostly trains these systems and the other mostly builds the sites and tools and teaches sailors to use them. That is the whole claim.
How to use it
| You are… | Start at… | Then… |
|---|---|---|
| New to all of it | Is this a job for GenAI? | The rest of Beginner, in order; you can use these tools well tomorrow |
| Using it daily and wondering why it does that | What an LLM is | The Intermediate tier, ending with Your documents at scale |
| Building helpers for your shop | Skills files | The Advanced tier, ending with Coding with models |
| Leading a shop | Working together | Where your text goes, then the verification checklist |
- As a course: one tier per evening. Beginner is self-contained; Intermediate explains the machine; Advanced builds with it.
- As a reference: the models table, the verification checklist and the glossary.
- As a handout: the cheat sheets and flyers.
Dates and drift
Facts about the vendors and the programs carry the date they were checked. These products change faster than a guide can; when the two disagree, the product in front of you is right and we would like to hear about it.
Go deeper
- How LLMs Actually Work: the under-the-hood book, if you want the mechanism.
- Training a Model: how a model is trained, evaluated and diagnosed, with weather models as the worked examples.
- AI/ML from the Ground Up: what machine learning is, from fitting a line to neural networks.
- The Basics on Latent Diffusion: how AI makes images.