You need a lot of context when decide the future. You need to look at many past structures zigzagging along the entails and jumping into translation doors. You don't "need" to leave the brain to make big discoveries and is sometimes/often infeasible. You just need more diverse data, possibly specific data too. AGIs can tell the humans what to implement, or what to do and gives results back.
So learning to walk/etc by short direct reward/context - not AGI much, there's a lot of such and you don't want to be doing every one to learn every thing. And non-mental R&D - not AGI so much, R&D to get results back is just gathering data! It's so inefficient actually. You just need more data. You'll need to do a lot of looking around your data to make discoveries (more quality data (high prediction probability = truth)). ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T99116eafd9a8e4b8-M24bcbed0899a214cfd54c56f Delivery options: https://agi.topicbox.com/groups/agi/subscription
