stevans@sri-unix (12/05/82)
I'd guess that of all of the computer paper consumed: %15 of the pages never have anything printed on them At least %15 of all printouts are never recieved by the intended party %50 of picked-up printouts are thrown away without any examination %75 of looked-at printouts have a maximum total of five lines read each Only about %5 of all printouts are kept for more than 6 hours No, these statistics have not been derived from my behaviour. Also please note that they represent the summed forms usage of all computers, not just high-tech sites such as most of ourselves. Does anyone have any comments? Please don't mail me hardcopy, though. Mark Stevans, U of Rochester, NY (allegra or seismo)!rochester!stevans
dwl (12/09/82)
Regarding the wasted printout::: The computer hardcopy I see most often is my paycheck! I usually only read it once. The bank probably reads it twice. It is probably thrown away by my employer. No, I don't mind receiving it! -Dave Levenson -BTL Holmdel
barmar (12/14/82)
Much has been said about the simulation of a hurricane not being a real hurricane, and the analogy to AI. I would like to answer David Sher's question about the differences. He hit it when he mentioned the possibility of connecting a huge fan as the output of the hurricane simulator. Both the "real" hurricane and the simulated hurricane have the same mathematics (assuming a perfect simulation), but we can tell the difference because the simulated hurricane is not uprooting trees in our "real" front yards. However, suppose the computer had the capability to make whatever effects it predicted true? This would include the feeling of wind on the face, the appearance of uprooted trees, etc. Would anyone doubt that it was a hurricane? This gets me back to AI. Consider one of the fundamental AI tasks, that of natural language recognition and generation. The nature of this problem already assumes that the computer will be interacting with the outside world, as opposed to just sitting there simulating intelligence. Indeed, if a human just sat there we would be likely to decide that it was not intelligent, since intelligence can only be recognized by noting the actions of the object. So, perhaps current AI programs are just "autistic", and the goal of AI research is to stop being AA (Artificial Autism) research. barmar