1962 D Penny Possible Day Day Out?

Whatever!!!!!!!!!!!!!!!!!!!!!!!!

No I haven't but I'll get right on it

Hey Spence if someone could come up with an app that they could put their coin on and it can read it like you read credit card they can make a lot of money it can get certified and priced out right there. I wouldn't think it would be any more difficult than reading a barcode

No expense I'm not heard of it. But it can have everything in the database and if the coin is worthless you just tells you just worthless. Is that what it does because all mine are worthless LOL

If you mean scanning a coin for errors, it would be enormously complicated, Chip. A QR code is a bit matrix that contains only ones and zeros as represented by black or white squares in a given location. The space of all possible matrices is relatively small, and all the machine is doing is entering a bit vector that then corresponds to some action like "go to this website." Scanning a coin would be a completely different matter. You would need to train a machine learning algorithm (probably some type of sophisticated neural net) on millions of images of coins that were all manually labeled as various errors and varieties or else as normal coins. You would also need tons of training data on damaged coins and worthless doubling so that your machine could distinguish these from errors and varieties. The problem is though that the space of possible damage is very large and machines are not good at extrapolating to inputs very different than the data they have already seen. Humans are very good at this, and (most of us) can easily see a coin with a minor scrape and another coin with enormous gashes that has been destroyed in a parking lot or with a belt sander and extrapolate that these fall into the same class. Machines have a very hard time with that because from the point of view of the image as a bit vector, the parking lot coin is as different from the coin with one scratch as the coin with one scratch is from a brockage error. Admittedly, I'm a theoretician and have very little experience with implementation of neural nets, but my impression from what I know of ML is that this would be very hard if not impossible with the current state of the art.

Speaking of which, why isn't Jethro Tull in the Hall of Fame?

They weren't allowed in the doors with that Locomotive Breath...