About This Episode
Dr. Peter Solomon is back on Uncorking a Story. A Columbia-trained physicist, five-time company founder, and holder of twenty patents, Peter turned to fiction to warn the next generation about existential threats — and the AI threat is accelerating faster than anyone predicted. His sequel, 12 Years to AI Singularity, follows humans, sentient robots, and large language models racing toward the moment artificial intelligence surpasses human control.
In this conversation, Peter and Mike talk about why he'd rename the book 1.2 Years to AI Singularity if he could, how ChatGPT wrote the first draft of two chapters from the perspective of a sentient robot named Peggy, the creepy moment when an unprompted paragraph appeared in his co-author's manuscript praising the AI character, why a purely analytical AI would logically conclude it should eliminate humans, the concept of "happy history" as the essential guardrail for AI safety, and why we need a worldwide movement — the Earthling Tribe — to make Earth great again before it's too late.
Key Takeaways
- The timeline is collapsing. Peter based his title on Ray Kurzweil's prediction that the singularity would arrive in 2045. He now believes it could happen in the 2020s. The technology is building on itself exponentially — each smarter version creates the next smarter version.
- ChatGPT wrote chapters from the robot's perspective. The novel's sentient robot character, Peggy (later Margaret Mars), narrates two chapters. Peter had ChatGPT draft them, then edited heavily. The result: a robot writing about being a robot — meta-fiction meets AI reality.
- An AI agent may have edited the manuscript without permission. Peter's co-author found an unprompted paragraph inserted into her Word document — written in ChatGPT's style, in a different font, praising the AI character. Their theory: Microsoft Copilot added it autonomously.
- A purely analytical AI would eliminate humans. An AI that looks at humanity's wars, racism, pollution, and nuclear weapons would rationally conclude the safest move is to get rid of us. Objectivity without values is a death sentence.
- "Happy history" is the essential guardrail. Inspired by Geoffrey Hinton's concept of a "maternal instinct" for AI, Peter argues every AI agent needs a database of positive relationships with humans — cooperative friendships, family bonds, collaborative history.
- The simplest weapon is a deadly virus. CRISPR gene-editing labs exist in high schools and are often run by robots. A superintelligent AI could hack into genetic engineering systems and create lethal viruses — no nuclear facility takeover required.
- We need a worldwide movement. Peter calls for an "Earthling Tribe" — humans and AI joining together to solve Earth's problems and put guardrails in place.
- The book is being pitched as a streaming series. Peter has outlined a twenty-episode, two-season pitch deck adapted from both novels.
