LiveForever could be a good idea?

Hi everybody,
I’m thinking to a project called “LiveForever”, it would be the creation of an avatar for people, let’s start to describe the project:

  1. Creation of a custom GPTs in which one person must upload all his biography, articles, ideas, etc.
  2. Upload short videos of his face while he is speaking, laughing, crying, etc.
  3. Upload short samples of his voice.
    When the GPTs ended we could have an avatar that can speak with us, using the face and the voice of the original one and using his “brain”. It could be interesting as a service for leaving a legacy after the death.
    What do you think? :slight_smile:

Well, it’s interesting, but with our current technology, it would be extremely hard. I recommend if it is to be used, using lip-syncing software as well as video generation.

Also, the GPTs would have to run for a long time to get the expected result, as EVERY SINGLE PERSON has their own personality and needs their own facial expressions and words, so you don’t just make a glorified Talking Tom.

I think as soon as the ChatGPT Plus subscription charges stop going through, the GPT personality is also dead.

You can try, but the AI won’t have pretraining or ability to browse to really be able to impersonate, and all the data uploaded is just retrieved in little snippets by a search feature. You could re-live the person’s life a dozen times over trying to make a training set of how the person would answer in response to a question, if there is any such written record.

Try a ChatGPT Plus GPT I made just for you (up for just a little while, for subscribers) ChatGPT - ZendayaAI and see what language AI can do someone with significant training knowledge.

thanks for your opinions, but my idea is simple, if I use many docs written by me, if I ingest a big doc with my biography, the first step could be to have a personal GPT that could answer using my ideas and experiences, then it could be improved with a text-to-speech for listening the answers with my voice. For the video could be another improvement but maybe it is complicated.

Doable. I can manifest this easily. This is a unavoidable future and we already have the tech now. Only its not integrated for the purpose. The problem would be to find funding. If you can find funding im willing to help in everyway needed as i share the same passion. [my email]

thank you for the support :wink: but I have had only an idea, I hope someone would develop that and I’m going to be happy to collaborate.