#deepmind quitting short works well for me at least: https://venturebeat.com/2021/06/09/deepm.. #reinforcementlearning is highly unlikely to get us towards #generalintelligence - it has always felt like there is something missing from the picture.