Image-gen 2.0 is awesome, but turns out Codex is better at turning a generated design into code if you prompt it to create a component map for the image first
Made this quick skill to do it in one step, see one shot Xcode screenshots without/with the skill
(link to skill in next tweet)
Here's the ui-gen skill. I'll try to add more of my own skills later. Hope that helps a fellow vibe coder
github.com/0xmts/public-skil…
Codex writes 100% of my code but I find it too chatty so I built myself a /progress [verbose | quiet] command
Minimal 1-2 line progress updates, refresh in place instead of taking over the screen
Surprised more people don’t know the CLI it’s open source. You can just do things
No ChatGPT? No problem! Just live next to an @OpenAI engineer who can come debug your random code irl. Thanks @romainhuet 😂
I was making a random iOS app with the help of ChatGPT but since it's down I decided to try to keep going with Bard... Let's just say it's not going great so far.
@romainhuet if you guys aren't back up soon I'm coming to knock on your door to make you write Swift for me.
I was making a random iOS app with the help of ChatGPT but since it's down I decided to try to keep going with Bard... Let's just say it's not going great so far.
@romainhuet if you guys aren't back up soon I'm coming to knock on your door to make you write Swift for me.
Siri and Alexa walked so GPT-4 could run and understand @romainhuet’s French accent live on stage 🤩
Happy @OpenAI dev day everyone!
Modern Pascal’s wager: it’s rational to be polite to ChatGPT just in case AGI goes rogue one day and turns against rude people first
brb I have to start a religion
The answer: memories
People keep asking me if we’ll be able to tell AI generated images apart from “real” photos. IMO absolutely not
The excitement of an unknown place, the taste of the food, locking eyes with someone new. Memories is the only thing we’ll keep making ourselves



















