ChatGPT + Computer Vision = A dictionary for what you see. By giving ChatGPT eyes, we allow it to answer questions about the world around us.

Details: 1. This uses a number of services to generate tags, recognize landmarks, and describe an image. 2. These outputs are massaged and injected into a prompt set up to answer questions using that CV output.

I did a rudimentary proof of concept of this a few weeks ago

Bro, imagine if this was integrated into some glasses for blind folks. I have a gentleman in my neighborhood that's blind, who I would come across once in a while, trying to cross the street when driving. I'd hop in my car and run to help the man. This is the next thing!

Love your tinkering.

Amazing work on this / @PatniShubh

This is an exciting prospect! By enabling ChatGPT to understand and interpret images, it can provide answers to questions about the world around us in a previously impossible way. The possibilities for this technology are truly limitless.

We're getting closer to the AI operating system featured in the movie "her " after chatting with GPT its response was that indeed it was technically possible to have a model like that , but we're still far away from developing a model capable of mimicking our emotional responses

Nice - can you share the t2s model/service you’re using?

It's Azure too

Dude! You are keeping opening eyes for the community, even the industry. Keep up the good work!