I recall once having a conversation with someone working on computer vision at Facebook, and they mentioned that the company had spent a considerable amount of talent & resources trying to get this feature to work cleanly and they did not get past the uncanny valley problems.
I second that, Uncanny Valley is exactly it. It was the same a few years ago on the web when web apps tried to duplicate native functionality (think filebrowsers etc.), it never worked right. In fact, Dropbox (however much I love it) still has that problem in their web UI.
Sort of the uncanny valley for intelligence. It astounds me how many people don't trust programs/automations when it is almost provably better at performing a given task. http://en.wikipedia.org/wiki/Uncanny_valley
but have we gotten past the uncanny valley yet? I've seen this attempted previously but it always seems very "off" in a way that engenders distrust. When we could just fix video streaming, it seems redundant to try to accurately replicate body language in a 3d avatar.
Interesting analysis. I didn't see enough variety of examples, specifically animations or even game AI.
Uncanny valleys was quite common in games and 3d back in the day. 3d modeling anatomy, like faces, bodies, etc., required a lot of training to pull off properly, otherwise you ended up with the uncanny valley. Now with motion capture, and 3d scanning (plus other tech of course) games and movies are basically humans acting, and there is almost no uncanny valley.
The video touches on part of what causes it, but the real cause is just attempting to recreate a human face REALISTICALLLY. Doing so in 3d was hard, and the industry avoided uncanny valleys in commercial content because of great artists, but software doesn't have the limitations of hardware. Recreating humans in hardware will take way more than a great artist, if it is ever possible. Chances are we will avoid the uncanny valley all-together, and just make smart robots that have that simplified cartoon look to them.
I'm not sure uncanny valley is the right term for this though I see the analogy.
Rather, I think it's yet another case where success (or at least impressive advances) in a narrow domain gets conflated with something much broader. Because if a human can do A they're clearly at least on the cusp of doing B and C too m
In the case of our digital personal assistants they've actually gotten pretty good at voice recognition, at least given certain parameters of accents etc. But what that means is that they're good at recognizing appropriate wizard incantations and taking the corresponding action. Get off script? It's worse than talking to one of those outsourced call centers we all hate.
Not sure what you mean by 'uncanny valley', this is just a simple case of raising expectations. The better designed the landing page, the higher the expectations of the product.
reply