Hey!

How to Train your Chatbot, my in-the-making book on using LLMs to build all sorts of cool stuff, just got another massive update. Since the last email, we have added three new chapters, encompassing over 30 new pages of content:

Chapter 8: Code Generation
Chapter 9: Agents
Chapter 11: Optimizing LLMs

So far, we have over 170 pages of practical advice for using LLMs in real applications, including:

Part 1: A finished 4-chapter part on how LLMs work and their capabilities and main limitations.
Part 2: Six chapters on prompt engineering and augmentation techniques, covering RAG, function calling, code generation, and agents, plus another chapter on optimization. The final chapter on fine-tuning will be covered soon.
Part 3: Three completed chapters on applications with full source code including a basic chatbot, a PDF Q&A bot, and an answering engine powered by search a la Perplexity AI.
Full source code for all the applications, including four more I haven't still written about but are fully implemented: a sales assistant bot with function calling and product generation, a data analyst bot that can generate and execute Python code to answer questions about any CSV file, a writing assistant and a coding helper a la Github Copilot.

Remember, getting early access to the book gives you all of the following:

Lifetime access to all book updates, plus special discounts for printed versions (if and whenever those are ready).
Access to fully functional Python code you can modify and reuse for your projects.
A private Discord channel where you can ask anything about the book and get help with the examples and your own code.

If this sounds enticing to you, here's an even better deal, a direct link to How to Train your Chatbot with a 50% discount.

How to Train your Chatbot - 50% off

And that’s all for today. Thank you for your continued support.

PS: Paid subscribers should receive a full discount coupon in the email version of this post. If that’s not the case, write to me.

New update for How to Train your Chatbot