The Waist Land

A certain technology company has a vision for the future of people hanging out together in cyberspace as embodied avatars. A lot of smart people are working on it, and a very large amount of money has been committed.

Yet in all of the descriptions that I have seen, the idea is to represent everybody only from the waist up. The reason, I believe, is that the technology is readily available to track peoples’ heads and hands, but not so much to track their feet.

I wonder whether this is going to be just a temporary glitch, or whether it will become enshrined as a standard. Perhaps from now on, this is how people will look in their on-line lives — they will not exist from the waist down.

If so, I am not sure that I am ok with this. For millions of years evolution has seen to it that we have a particular arrangement of brain and body. The human brain has evolved to be highly attuned to that human body.

Maybe we shouldn’t just throw out half of the human body because of a temporary technical inconvenience. We don’t want our future reality to be an impoverished Waist Land.

Sigh. November is the cruelest month.

3 thoughts on “The Waist Land”

  1. It would surprise me if based on the motion of the upper body (integrated across a longer time interval) the motion of the lower body couldn’t be predicted plausibly (at least for a viewer who doesn’t really know what’s really going on). So I think there is no need to track the lower body, the reason why it is missing right now is that they just skipped on this problem because the technology of that certain company is very new, they worked on the problems they considered most important at first.

  2. We’ve done quite a bit of research in this area at our lab, and it turns out that it’s an intractable problem, because there is insufficient information. You can’t tell from movement of the upper body what the feet will do. For example, when you first start to walk from a standing position, you kick out your first foot before your upper body moves, and then the rest of your body follows afterward.

    My big worry about their approach is that they are quite ok with replacing your actual body language and facial expression with a Machine Learning proxy when it is convenient. And that can create a false sense to other people of your intended emotion and purpose as you communicate. I consider that sort of error to be a disaster, from a human perspective.

  3. Right, I didn’t think of that. Also, I guess the way networked games handle latency by playing with animation speed and blending in connecting motions is not good enough in these scenarios. Then indeed you have serious problems. But I guess they will solved by making the people themselves adapt to the new technological realities. After all, we’ve adapted to 200ms latency mice also, not expecting too much body language of a VR conversation will be the norm. 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *