This is outdated.
|Latest docs (12.08.21)|
|Project timeline and objectives||https://github.com/semiosis/pen.el/tree/master/docs|
|Video demo of
||Augment Minds 2021: Demo of Loom and Pen.el - YouTube|
|Bash prompting README||http://github.com/semiosis/pen.el/blob/master/docs/using-prompt-functions-in-bash.org|
Prompt-Engineering Part 1: Building an environment
Summary of talk
I received a key for the
OpenAI API 2 weeks ago, just a couple
of days before my birthday.
I had applied around 5 to 10 times in the last year and I had finally been given a key.
I decided since the hour is late I would work on productivity tools that anyone can use, to facilitate programming in the new paradigm.
pen.el : Prompt engineering in emacs
Prompt-Engineer Mode is a global minor mode for emacs that facilitates the creation, development, discovery and usage of prompts in your emacs.
- Create elisp functions based on GPT-3 prompts
- Chain GPT-3 queries together using keyboard macros and functions
- Interactively query, generate and transfrom both prose and code
- Use GPT-3 as a search engine within emacs
- Search the internet
- Search documents
- OpenAI API (GPT-3)
- Requires tensorflow 1
- DeepAI API
- OpenAI API (GPT-4)
Usage / Workflow
prompt-engineerer-modeglobally (on all buffers)
- Have your
.promptfile open in one window
- Have the document open you want to apply prompt functions to
- This could be source code, a chat window, or any kind of text document.
- Apply functions to do all sorts of things
- Generation of prose
- Generation of code
- #029 GPT-3, Prompt Engineering, Trading, AI Alignment, Intelligence - YouTube
If this article appears incomplete, it may be intentional. Try prompting for a continuation.