https://github.com/transitive-bullshit/severance-wellness-memes
Roast your Twitter profile with AI-generated Severance wellness memes.
https://github.com/transitive-bullshit/severance-wellness-memes
llms memes severance twitter
Last synced: 11 months ago
JSON representation
Roast your Twitter profile with AI-generated Severance wellness memes.
- Host: GitHub
- URL: https://github.com/transitive-bullshit/severance-wellness-memes
- Owner: transitive-bullshit
- License: mit
- Created: 2025-02-18T15:53:27.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-16T09:44:12.000Z (over 1 year ago)
- Last Synced: 2025-03-16T10:27:53.283Z (over 1 year ago)
- Topics: llms, memes, severance, twitter
- Language: TypeScript
- Homepage: https://severance-wellness-session.com
- Size: 5.57 MB
- Stars: 9
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- Funding: .github/funding.yml
- License: license
Awesome Lists containing this project
README
# Severance Wellness Memes
> Roast your Twitter profile with AI-generated Severance wellness memes.
- [Why?](#why)
- [Examples](#examples)
- [Feedback](#feedback)
- [TODO](#todo)
- [How it works](#how-it-works)
- [Tech stack](#tech-stack)
- [License](#license)
## Why?
The work is mysterious and important.
Praise Kier.
lol but seriously though, with [the release of GPT-4.5](https://openai.com/index/introducing-gpt-4-5/), I was really curious to see how good the latest LLMs were at generating subtle, personalized humor.
After a lot of iterating and testing with every frontier LLM, **the results were... mixed... except for GPT-4.5**, which was able to generate some genuinely funny and surprisingly nuanced jokes. In my experience, \`gpt-4.5-preview\` generates really solid results about 50% of the time, so in a batch of 10 memes, there's generally a few standouts.
## Examples
## Feedback
Feel free to send feedback by replying publicly to this [Twitter/X thread](https://x.com/transitive_bs/status/1898379925155156203). You can also [DM me](https://x.com/transitive_bs), but I am more likely to reply if you are following me and the conversation is public.
I prefer public convos so the answers can help other people who might have similar questions, but if it's related to billing, please [DM me](https://x.com/transitive_bs).
## TODO
- Post-MVP
- [ ] add sitemap
- [ ] improve landing page hero section
- [ ] improve header section + mobile support
- [ ] add basic pricing page w/ explanation for why I had to add the paywall
- [ ] add url resolving to `resolveTwitterUser` to give more context
- [ ] show generated date and model on fact detail page
- [ ] consider switching from prisma to drizzle
- [ ] consider switching from supabase postgres
- [ ] allow paying for more tweets + generations
- [ ] track third-party fact url access and image access for stats
## How it works
The workflow to generate memes for a user:
- Fetch their twitter profile and 200 of their most recent tweets.
- Use \`gpt-4.5-preview\` to generate a list of text-based wellness facts based on a [long prompt](./src/lib/generate-wellness-facts.ts), some examples to guide the model, and the user's unfurled tweets as an array of plain text.
- Use \`gpt-4o-mini\` to try and [extract the user's full first and last name](./src/lib/extract-user-full-name.ts), so we can call them by their Severed Lumon name (eg, Helly R).
- Use [Vercel's dynamic OG image generation](https://vercel.com/docs/functions/og-image-generation#getting-started) (which is powered by [Satori](https://github.com/vercel/satori) under the hood) to generate the severance meme images on-the-fly.
### Tech stack
- [TypeScript](https://www.typescriptlang.org)
- [Next.js](https://nextjs.org)
- [Vercel](https://vercel.com)
- [OpenAI](https://openai.com)
- [Postgres](https://www.postgresql.org)
- [Prisma](https://www.prisma.io)
- [Stripe](https://stripe.com)
- [SocialDataTools](https://socialdata.tools)
- [Agentic](https://github.com/transitive-bullshit/agentic)
## License
MIT © [Travis Fischer](https://x.com/transitive_bs)