I hate expense reports. I hate them so much, I let things pile up. But that just makes things worse.
Yesterday, I reached my limit. In an act of desperation, I dragged a folder full of receipts into ChatGPT and asked it to create the report. A minute later, I had a finished expense report, downloadable as a .CSV.
What this essay explores
- Understand how context windows and training data shape an LLM’s long-term and working memory.
- Use RAG to bring fresh company data into an AI product’s context window.
- Recognize why LLMs need product functionality for arithmetic and other computer-native tasks.
- Make your product’s functionality available to LLMs through function calling.
- Distinguish practical tool use today from the broader promise of fully realized agents.
- See how function calling can help analyze spreadsheets, surveys, support messages, and user reviews.
Growing context windows make it possible for AI to hold more of a product’s state and workflow at once.
Key takeaway
As a product builder, you need to stay ahead of the AI curve. In this post, I discuss where AI is going and how to keep your product at the frontier.
