I think this comes from the theory of general artificial intelligence where your AI would have the ability for self improving. Hence it could develop any capabi

I think this comes from the theory of general artificial intelligence where your AI would have the ability for self improving. Hence it could develop any capability given time and incentive for it. There are interesting videos on the subject on Robert Miles channel on AI safety: https://www.youtube.com/channel/UCLB7AzTwc6VFZrBsO2ucBMg https://www.youtube.com/channel/UCLB7AzTwc6VFZrBsO2ucBMg