The way we work is we "do" predict the next word based on context, right away
without "extra" thinking, it happens 24/7 nonstop, decisions get made always.
What we do is, if we are not confident in the prediction, we will not predict
it, but rather something else.... but we will be confident in predicting the
context (question) still, as it is not answered yet. We collect data at likely
good domains and come back to the question. I mean we do predict the answer,
but we also predict that it isn't the right answer. Based on this context, we
predict what data to now collect.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink:
https://agi.topicbox.com/groups/agi/T15fa07153b838749-Mf797c46b6931dfe3b4dd20a6
Delivery options: https://agi.topicbox.com/groups/agi/subscription
