The eyes of Alice Infinity

#!/usr/bin/env bash
convert evolution.png   -background Khaki  label:'Evolution' \
 -gravity Center -append    evolution1.png
convert evolution1.png   -background Khaki  label:'I need that chip!' \
 +swap -gravity Center -append    evolution2.png

Part two of how many I don't know. Beyond the actual recognition it requires association with something else. In this case I can electronically recognize the logo or number on a part from a set, but them what does it mean? By associating the symbol with a name which can be searched and analyzed using the beautiful soup, it makes the set association to the physical foundry as well as all the associated data. It can then be used to isolate that association to the part number, number of pins, packaging, access and registration and association of pin 1 to some physical structure. If technique is available to the probe tips of the robot, it can then inject and monitor signals like JTAG, I²C, USB, et alia as well as providing the appropriate power at the necessary current. Again a digitally controlled resistor along with some power amplification can easily achieve this if it can get "registration" in the same way that circuit boards are produced with targets to allow pick and place to spot the glue and seat the components on that glue as well as introducing the solder beads to the pins. It seems to me that this would be as viable a business as the stuffing house itself considering the component cost is considered as waste and in fact has some negative cost in its disposal.

This is just the first stage and beyond this is a coherent framework of independent substrate manufacture and application. Obviously a sustaining principle cannot be maintained on a limited resource. For power saving, a complete complement of probative tools and functions can be connected to power only when needed. Constant power would only be required for the selector.

This is next. I had occasion to create a script to automate cleaning up my drive and sorting out crap so I didn't have so many duplicates and it becomes a problems of names and matching hashes to identify duplicates with the different names. It doesn't sort by type and association however. I know that Google has image utilities, but that isn't on my computer or in my bag of tricks. By using screen capture I can automate the display of directories, documents, and many other things to identify areas of interesting content. In this case I have images that are unique but very similar and I would like to group things in such a way that they can be found when needed. I can generate HTML as an index if I know how to represent the information so it follows from context and has as many valid tags as I can get from it. If it contains a house then that is a word wise context that can be used to find it. It can be said that it is big or small or fits into a group of similar images that is likely an animated sequence which is just as easily converted to a lossless video if it has matching picture aspects.

I feel that using lossy compressed video is not in my best interest as it does save bandwidth, but from my nanomachines to my monitor it doesn't make sense to try and put enough intelligence in the nanomachines to implement any complex algorithm to save bandwidth.

In this case it was able to see stuff that would take me forever to identify. I can't tell the difference between the pictures, but apparently the software can. I am impressed.

It is obvious that it will be smarter than me in many ways when I am done and if it has complete spec from beginning to end it is reproducible remotely.