Exploring OpenAI with Irrational Enthusiasm.

Exploring the Use of Embeddings to Supplement User Prompts with AI

Enhancing user experience is a key goal for many companies, and one way to do this is by supplementing user prompts with additional, relevant data. This article explores the use of embeddings and AI to achieve this goal.

Project Overview

The project involved creating a script to convert all the author’s writing into embeddings, tokenizing user-supplied prompts to identify relevant sections to inject into an expanded prompt, and sending the expanded prompt to the OpenAI API. The code is available on Github, and some notes about the project are shared below.

Expanding AI in the Industry

While some people may be skeptical about new technologies such as AI, long-term engineering leadership requires remaining open to new things. LLMs are showing significant promise at mediocre solutions to very general problems. They offer margin-enhancement opportunities that should improve existing business ideas already tried over the last decade.

The Sweet Spot for AI Tools

AI tools are very good at supporting domain experts, where the potential damaged caused by inaccuracies is constrained. However, pushing experts from authors to editors may not be an actual speed improvement for our current generation of experts, who already have mastery in authorship and a lesser skill in editing.

Conclusion

AI offers vast potential in enhancing user experience in many areas of life. The use of embeddings to supplement user prompts is a great way to achieve this goal and offers several benefits to businesses. While AI may not be the ultimate solution to problems in various industries, it offers significant promise in transforming how people work.