Oracle Content Management - AI writing assistant component! WIP…

Recently my twitter feed has been full of GPT-3 posts using language models that use deep learning to produce human-like text. From summarising content on a page to producing well formatted answers. Here is an example from a friend who's been trying it out on https://beta.openai.com/playground

Now... I know that I would not have been able to provide as clear of an answer to that question..

With that said wouldn't it be great if we could take GPT-3 and use it with Oracle Content Management to do some of the following:

  • Rewriting your content in different styles for example enhancing and improving the grammar or modifying the content and condensing it to fit a character limit or making the copy sound more positive.
  • Maybe you need someone to assist and provide personalised ideas and content suggestions on topics.
  • Writers block? The AI can continue writing content for you to review and modify.
  • Provide feedback on your content and provide guidance and tips on improving the content and structure.
  • Improve and auto generate SEO Keywords and descriptions based on content held within a page
  • Autogenerate RSS feeds with punchy engaging headlines.

I started playing with a concept last night and creating a basic integration with OCM; experimenting with some of the ideas above.. And now I'm scared!

It's too good.. It's still very early but I'm hoping to get a sample out soon before Christmas for you all to try as an OCM large text field component you can install and apply!


If you are looking to play with GPT-3 head over to https://beta.openai.com and get a key and checkout some of the new samples ie here is one connected with Hacker News to summarise top stories.

GitHub - jiggy-ai/hn_summary: Summarizes top stories from Hacker News using a large language model and post them to a Telegram channel.
Summarizes top stories from Hacker News using a large language model and post them to a Telegram channel. - GitHub - jiggy-ai/hn_summary: Summarizes top stories from Hacker News using a large langu...