Sam Casey retweeted
Introducing Chat Collect 📥
Easily collect emails from your GPT users.
Start boosting engagement with your GPT users natively from the Chat UI in preparation of the GPT store revenue sharing program.
No code required.
Sam Casey retweeted
Friends, don't be a grinch!
Head over to Product Hunt and give RingleDingle some love, and spread the holiday cheer! 🎄🔔🎅
@Sprintfolio_co designers in our ai-powered design school have inspired some really cool features and update.
launched RingleDingle on Product Hunt today
right now we're in 18th place... meh
hoping to get top 10 by the end of the day!
link to upvote below :)
producthunt.com/posts/ringle…
thanks in advance if ya vote!
launched a new feature on RingleDingle yesterday
you can now:
- edit song lyrics
- regenerate the card's song
- regenerate artwork in a new style
pretty happy with how this turned out
now time to watch Posthog and see if the ppl who requested this actually use it lol
🎶 Introducing RingleDingle!
RingleDingle lets you create custom musical e-cards to delight your friends and family. No artistic ability required!
Built this as a fun little side project last weekend with @DroopyCodes
Gonna keep improving it and launch on product hunt soon for the holiday season
Any feedback is much appreciated!
one of the best parts about GPTs is that they’re a great way to test tools you’ll use for an agent before creating your own UI
build a tool, expose it via a rest api, test it with a gpt, test it and see what custom UX would make it better, and then build your agent
for the past few weeks i've building an AI agent from scratch it's been a disaster
i've used 3 different approaches, and i'm finally giving up and just going to use the new Assistants API from OpenAI
on learning #2:
for tasks that take a long time, chat UX is disorienting
it feels weird to leave a chat and then come back to the latest message
this isn't good UI yet, but some progress tracker towards the larger goal feels necessary
you know Twitter is about to be littered with some fire demos when you see the “we’re really busy sorry for delays” warning on ChatGPT
building generalized products is way more fun than targeted ones from a tech perspective, but way less fun from business perspective since no one will want what you’re building
i’ve reached a plateau in my speed of building with llms
need to learn more about testing and benchmarking to get faster
there’s gotta to be a faster method than running py main.py a hundred times a day and reading all the logs





