GPT-3 in Production
Published by O'Reilly Media, Inc.
Including language models in your existing apps
If you don’t have an internal team to train your own machine learning models, you can still be on the cutting edge, using the most advanced AI models, by leveraging the powerful OpenAI API in your production apps. OpenAI’s models, GPT-3 and the newer Codex, which promise to simplify the programmer’s job, have generated a lot of buzz in the last couple of years.
Join expert Bram Adams to learn what all the excitement is about. First, you’ll examine the business value of “stochastic” programming and how you can include natural language in your applications to provide more value to your clients or customers. Then, you’ll get acquainted with GPT-3 and Codex and examine use cases to integrate into a wide array of day-to-day problems your business may face. Follow along while Bram live-codes two toy apps and shows you how to access data and create new data to be parsed by programs using the API in JS and Python.
Note: You can code the apps along with Bram if you sign up for access to OpenAI; you're also welcome to just follow the demonstrations.
What you’ll learn and how you can apply it
By the end of this live online course, you’ll understand:
- How to use the semantic search API to use natural language to find exact things
- How to use the completion API to create parseable lists from natural language
- How to use Codex to increase the impact of your organization’s codebase
- The questions to ask about your app regarding when to include the OpenAI API
- How to use semantic search to surface relevant data
- How to use the completion API to get consistent results
And you’ll be able to:
- Explain to your team what GPT-3 and Codex are, and when to use them
- Know where GPT-3 fits into your app
- Think differently about how code fits into your business
This live event is for you because...
- You’re a software engineer or technical leader who is excited to add new products and services to your business.
- You work with people, and machines that interface with people.
- You want to become a successful operator of probabilistic technology.
Prerequisites
- A working knowledge of NodeJS and Python servers and HTTP calls
Recommended preparation:
- Explore OpenAI API (website)
- Read “Introducing GPT-3 and the OpenAI API” (chapter 1 in Exploring GPT-3)
- Read “GPT-3 Powers the Next Generation of Apps” (OpenAI blog post)
- Read “GPT-3: Your New Lifelong Companion” (blog post)
Recommended follow-up:
- Finish reading Exploring GPT-3 (book)
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Introduction (15 minutes)
- Presentation: Integrating natural language into your current apps in production for fun and profit; the ease of use of OpenAI’s GPT-3 and Codex, two of the world’s most powerful ML models
The completion API (50 minutes)
- Presentation: The completion API; pros and cons; tweakable values; API wrappers for your favorite language
- Q&A
- Break
Live review (15 minutes)
- Demonstrations: Using the OpenAI Python library to call GPT-3; prompt training a bot with reasoning, wit, and tabular data
The search API (50 minutes)
- Presentation: The search API; pros and cons; how to make your data semantic
- Q&A
- Break
Live review (15 minutes)
- Demonstrations: Searching different types of lists at the same time; classification and answers
Codex model (50 minutes)
- Presentation: New models; code is everywhere; the purpose of code at organizations; Copilot; What I’ve learned building Stenography
- Q&A
- Break
Live review (15 minutes)
- Demonstration: Making a simple web app with natural language
Wrap-up and Q&A (30 minutes)
Your Instructor
Bram Adams
Bram Adams is the founder of Stenography, a GPT-3/Codex-based API that translates code into natural language. He also makes art using artificial intelligence and augmented reality frameworks and libraries.