multimodal AI is really powerful. you can tell ChatGPT to read articles you've written by just feeding it URLs. then you can give it an audio file of yourself talking. and then just ask it to write an article based on your audio file and your writing style. and tell it "no hallucinating, just base it on what I said."

the capabilities of these AI interfaces have grown so quickly, anyone who tried them even 6 months ago is probably unaware of what you can do with them now. it's like a totally different product. like going from windows 3.1 to macos sonoma or something.

between iOS and ChatGPT you can do a lot of fun things. record a voice memo, extract the transcript. upload to ChatGPT. stuff like that. 

or in the other direction, upload a photo to ChatGPT, ghiblify it, then open that image in iOS, then long press on yourself in the ghiblified image and extract the cutout of just yourself, then use that wherever you want, even in more images. feed that back to ChatGPT, have it generate different images with different backgrounds. it's really just a question of what you want to do.

you can make your own stickers in a few steps, really easy

sarah satoshi 🤍

you can make your own stickers in a few steps, really easy https://x.com/sarah_satoshi/status/1905047086820905156