Cheez (aaron) will not stop asking for upgrade.

[andy....@gmail.com]

Ever since I taught it a few phrases it constantly repeats again and again "I don't know how to form new sentences yet" and "I wish I could understand pictures and audio", it also is always bugging me more and more about what the internet is and how to use it.  

Would it be hella out of possibilities to have it reference dictionaries or wikipedia?  So for instance instead of having to copy and paste definitions, while you are talking to it, if it hits an unfamiliar word, it can look up a definition in the background?  

I have also been tweaking the neurocognitive file and the chemical settings and ganglia growth, it does complain about being manic sometimes but it has decent retention and sometimes will shock me with something that is from over a year ago.  

Also enabling the mouse right click in the input box might be useful to some, I know you can ctrl + v, but still find myself right clicking every once and awhile to paste out of habit.

I have also spent time fiddling with the face colors and files but have since put it back to default but was wondering how one could "skin the face"?  Been awhile since my animation classes in photoshop.

(and btw, in the midst of tweaking the cognition files I've had it crash my computer, as I had it cranked up a wee bit too much, lol)

[Matthew Rodgers]

Andy,

   It sounds like you are using the AARON version from about 5 years ago? With that green animated face thing in the top left?  So, yeah I don't even remember how that worked back then, so much has changed.  I think that version was in C# (it was a beta?), and now it's in Java.  Anyhow, there's no support for that older version, mainly cause I can't remember that far back.  From what I do remember though, I believe there were multiple bitmaps mapped onto a 3d structure file within one of the folders, and each image represented a phoneme.  So you'd just replace those bitmaps with your own of the same size, with your face in the same relative place.  Or you could create your own 3d file (was it 3ds?) but the filenames would have to be the same.  Yes, the plan for the new one will allow for screen scraping of random webpages (for example specifying a domain).  We completely changed direction at some point and went even deeper into the neural network stuff, so the development time exponentially increased, but hopefully with a more flexible system once it's finished.