Running list of interesting ideas / possibilities that @OpenAI's GPT-3 could open up →

(1/n)
Disclaimer – this is more of an exploration of things that could be useful & possible. Actual feasibility might be a bit iffy.

I do not (yet) have access to GPT-3 (if somebody can help get me one – please do – I really want to build some of the ideas listed here!) 👇🏼
1. Wireframes to Mockup.

Input a simple quick wireframe and get a final polished mockup as the output.
– Ability to add Visual Cues
– Icons using Iconscout / Fontawesome APIs
– Images using Unsplash / Shutterstock APIs
2. Headline Generator

Input seed text. The tool outputs different variations of headlines generated from that.

You could choose:
– Level of SEO friendliness
– Simplicity of English
– Translation
– Targeted to X Demographic
3. PRD to Mockups

Input a PRD / writeup of the functionality of the app, and the tool outputs a Figma mockup.

You could add:
– Inspiration Websites
– Visual Prompts
– UI Components
– CTA weights
4. Supercharged Lorem Ipsum Text

Never fuss over writing copy again. Input the Designs with seed text – and then the tool adds in the text.

You could choose:
– English proficiency
– Languages
– Target Audience
– Industry Segment
Although, GPT-3 is a Language Model, it could still work integrate well with @figmadesign @sketch @Photoshop etc, since they use a code schema to store the graphical representation of the UI / Design. Which is (sort of) like the language of these software.
5. Book / Blog Reviews

Train the model on self reviews of different books / blogs you've personally read.

If you want to read a new book – run the model first, read the review, and then decide if it is worth reading it.
You can follow @dhvanilp.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled:

By continuing to use the site, you are consenting to the use of cookies as explained in our Cookie Policy to improve your experience.