Do you have a dumb question that you're kind of embarrassed to ask in the main thread? Is there something you're just not sure about?
This is your opportunity to ask questions. No question too simple or too silly.
Culture war topics are accepted, and proposals for a better intro post are appreciated.
Jump in the discussion.
No email address required.
Notes -
We had a death in the family recently. It will be a simple funeral - a viewing, a cremation, and a burial of her ashes. I chuckled a bit when the funeral home asked us for a DVD for the photo slideshow during the ceremony. It got me thinking - how is AI technology and AR/VR going to change the future of the funeral industry?
Imagine that an AI avatar was trained on voice recordings, videos, photos, and text of the deceased. You visit the cemetery with your family and you all don the VR goggles, stepping into the living room of grandma's house. She's tending to the garden and her avatar ad-libs about her tomatoes and the recent weather. Just as you remembered from a few years ago before she had to go into the nursing home.
If you've seen the incredible improvements in image and video generation in the last 2-3 years, as well as the improvements in text-to-speech (see a previous Friday Fun thread post that I shared) you'll probably agree with me that this is something we'll see in our lifetime. Yes, we'll have a period of uncanny valley, but when it's fully ironed out, there will be a convincing digital copy of ourselves floating in the ether.
Everyone spends some time chatting with grandma then she excuses herself to take the cookies out of the oven. You decide it's about time to grab lunch with the family and say goodbye for now.
The funeral home charges you for the disposable insert in the VR goggles that soaks up your tears.
A lifetime? More like 2 years. The only thing we can't do today (with good results) is the interactive AI avatar or real time video.
The other reply is right, you don’t have anywhere near enough data on Grandma (unless she’s a famous celebrity with a 40+ year TV career) to train an AI that ‘feels’ like it’s her. She’ll make an infinity of hand movements, facial tics, use certain words and the illusion will be instantly destroyed.
It’s as much about learning the things she would never say as it is appearing like an authentic old person. For it to be fully authentic those Meta glasses will have to catch on so we can record dozens of hours of mundane interactions with others. I would estimate that a few thousand hours would be enough for a good facsimile.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link