> I saw a picture of a guy that famously looked like Ted Danson. He was just some guy somewhere that happened to look a lot like him, and I forget whether I saw his picture online somewhere, or if they showed his picture on a late night talk show or something, but in any case, I saw the guy's picture, and INSTANTLY, two things registered to me: (1) Whoa, this guy looks like Ted Danson, and (2) This guy isn't Ted Danson.
>
> The human mind is amazing enough in its ability to register thought #1, but I'm at a total loss for how or why #2 came to me. I didn't know what looked different enough from the actual Ted Danson that I was able to identify that that wasn't him and was just someone that looked a lot like him, but in a split second my mind was able to interpret the image in front of my eyes and figure all that out.

I totally agree that this is a truly amazing skill. However, this only convinces me that I would not be able to create (or it would be vastly harder to create) a robodog that would convince you that it was *your* dog, i.e. make it as a replacement for your current pet. It says little about the ability to fool you into thinking that it was *a* dog, and not some specific dog you were already familiar with. One is a matter of specifics, the other is an abstract concept. We recognize poodles and St. Bernards as being "dogs", even though we would never confuse one for the other. And we can certainly tell the difference between one poodle and another, especially if we were intimately familiar with one of them to begin with. But I still think it says nothing about our ability to distinguish a meatdog from a robodog. I could be wrong. BUT I AM RARELY WRONG! RAAAR! ;-)

>
> Of course face recognition is a skill some people are better at than others. But there are other ways we recognize each other, another being voice recognition. If memory serves, you're not actually very good at face recognition, but you have a spectacular sense of voice recognition.

I wouldn't call it spectacular. But I do recall one time I met an old high school classmate several years after we had graduated, and I was at a loss for who he was when he came up to me. He said "You don't know who I am, do you?" and I said "Keep talking" and after a few more words I remembered. So, yeah, I would have to say my voice recognition is better than my face recognition.

>A really big one that we're less conscious of is the manner of our body movement. People walk differently, gesture differently, and we pick all that up, knowingly or not. Here's a wild one: we have a great sense of *height* recognition. I saw an interview with a retired agent of some kind, FBI maybe, and he was saying that the best way to disappear into a crowd if people are chasing you is to walk on your toes or stoop down just a little bit. Because apparently the first criterion our brains use to pick someone out of a crowd is height -- and the difference of just two inches or so can be enough for our brains to subconsciously filter someone out as a possibility.

This *is* pretty wild. I'll have to remember this one. :-)

>
> Now, obviously this is people, not dogs, and it's talking about *individual* people, not people as a biological species. The point is just that our brains are SO tuned to recognizing things like ourselves. Would I be able to identify trees or airplanes or mountains with as much precision? Not even close. But dogs have our most fundamental biological traits in common with us. They're anatomically analogous, they're emotional in the purest ways we also are, they're expressive, and they're personable. Try to simulate all that in a robot, and I think those amazing parts of our brains that discern one voice from another, one face from another, one walk from another, would start screaming at us that something is off.

This is certainly possible. However, I'm not convinced. Part of the problem is that our internal "facial recognition" pattern matching ability is very prone to see matches where there are none. We see faces EVERYWHERE--in clouds, on the moon, in oil stains in the driveway, on tortilla chips. Heck, don't forget about the Old Man. A collection of rocks and cliffs that vaugely resembled a human profile, beloved by thousands of people to the point that many mourned his "passing" much the same way as if a friend had passed on. ROCKS inspired that emotion. Rocks that SORT OF VAUGELY looked like a human face. Sure, if you believe in the "Uncanny Valley" effect, you can explain it. But I don't think the Uncanny Valley plays into this discussion much, beause we're assuming that we're beyond that valley and very far up the next uphill slope with our proposed robodogs, since we're assuming that they are pretty much indistinguishable from real dogs.

Anyway, I agree with you that the human brain is amazing. I disagree with what seems to be your underlying feeling that it (or at least, less amazing animal brains) are too complex to ever simulate or duplicate.

And yeah, I know I said I was done after that other post. But I forgot I hadn't responded to this one yet. So :-P

-- Dave