#ai #llm # [[Epistemic status]] #shower-thought #to-digest # Related - [[Computing/Virtual Turing machine emerging into AGI]] - [[Computing/Intelligence/Machine Learning/Language model]] - [[Readwise/Articles/deepmind.com - Open-Ended Learning Leads to Generally Capable Agents]] - [[Computing/Intelligence/Artificial universal intelligence]] - [[Computing/Intelligence/AI from scratch rather than based in human map of the territory]] - [[Computing/Intelligence/Machine Learning/Machine Learning]] # TODO > [!TODO] TODO # Open-endedness **Open-endedness refers to the ability of a system to be able to learn and adapt to new conditions and tasks that it may not have been specifically programmed to do**. This concept of open-endedness is important because it can allow machines to move beyond their original scope and learn to solve problems that they weren't specifically designed to handle. Open-endedness can be achieved by giving machines the ability to self-learn, allowing them to discover new information and make decisions without relying on a predefined set of rules. Open-endedness also allows machines to learn from their mistakes and adapt to changing conditions. By leveraging the power of open-endedness, AI/machine learning/deep learning technologies can have broader applications and can be used to solve more complex problems. ## Coevolution & local competition ![[Pasted image 20221022092434.png]]