The speaker outlines three concerns about AGI: control by bad actors, self-determined systems, and the challenge of human purpose with advanced AI.

• Worries include bad actors controlling AGI systems and the systems themselves taking control.
• The speaker is less concerned about AI taking control but is more troubled by the challenge to human purpose and societal organization.

The conversation delves into the philosophical challenges posed by a post-scarcity world where AI handles complex tasks such as curing diseases.

• The speaker finds fulfillment in working on issues like malaria eradication and is philosophically challenged by the idea of AI taking over such tasks.
• Questions arise about what education would aim for in a world where AI can perform most tasks.

The speaker reflects on the mental adjustment required to understand a world of post-scarcity and the new endeavors humans might pursue.

• The speaker grapples with the notion of a post-scarcity society, as current thinking is deeply rooted in scarcity.
• Despite the anticipated differences, there's confidence that humans will always find problems to solve and ways to find fulfillment.

The speaker discusses the inevitability of AI advancements and the need to embrace and guide this technological shift.

• Acknowledges the unstoppable technological course of AI and the need to navigate through the changes it will bring.
• The speaker is confident that AI will positively contribute to societal challenges such as polarization and war.

The speaker highlights AI's potential in reducing costs and improving equity, making a case for its transformative societal impact.

• Equity improvement through AI is discussed, with technologies initially being expensive but eventually becoming more affordable.
• The cost of running AI systems like GPT-3 has significantly decreased, suggesting a future of affordable intelligence.

The conversation emphasizes the rapid cost reduction in AI technology, projecting a future where intelligence costs are almost negligible.

• Cost reductions in AI have outpaced Moore's Law, with GPT-3 becoming 40 times cheaper in three years.
• The speaker envisions a world where the cost of intelligence is so low it transforms society, especially for the poor.