Robert Wright views AI as a threshold event in planetary history, where machine learning mimics natural selection by reverse-engineering cognitive functions. He argues that humanity needs a moral upgrade, including overcoming cognitive biases and fostering global cooperation, to safely manage AI's risks, such as job loss, bioweapons, and cyber hacking. Wright warns that without deliberate intervention, market-driven AI development will likely lead to negative outcomes, including sycophantic AIs and a worsening meaning crisis. He sees evidence of the singularity dynamic in coding agents improving AI models and notes that the time for AI to complete human tasks is doubling every seven months. Wright concludes that human superintelligence already exists through collective intelligence, which AI can enhance, and recounts Ed Fredkin's optimistic view that superintelligence would treat humans well.