Suppose my feeling that complexity - complicatedness - is the issue that makes AGI so elusive turns out to be right. Then, someone should be able to show that a simple model actually can work - as long as it is kept really simple. Next he would have to be able to show that the model which can be applied to one 'field' (in this case one IO modality) could be applicable to another 'field' with some adjustments to deal with the other kinds of IO events. Finally he should be able to show that they can be combined so that it can use learning from one field and apply it to another. Finally he has to show that the same can be done with other 'fields'. The only issue is that everything would have to be kept at an extremely simple level and that 'genuine learning' is actually taking place. Defining and developing general learning is the only problem other than complexity. Jim Bromer
------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T187cd9f14076b86f-M898f00104b1fb37d3b5b8072 Delivery options: https://agi.topicbox.com/groups/agi/subscription
