What if the best way to explore the galaxy is to launch one ship that builds copies of itself? This episode unpacks von Neumann probes, from John von Neumann’s insight that exponential replication beats launching fleets, to Robert Freitas’s 1980 engineering study that adapted the Daedalus starship with a 443 ton seed factory spending 500 years building its industrial base. We also cover Olivia Borgue and Andreas Hein’s more feasible 70 percent self-replicating design that brings microchips from home, and Anders Sandberg and Stuart Armstrong’s scenario of dismantling Mercury to build a Dyson swarm and launch probes in 32 years.

At 10 percent light speed such machines could fill the Milky Way in about half a million years, which is why Frank Tipler argued in 1981 that their absence proves we are alone. Carl Sagan and William Newman countered that unrestricted replication would consume the galaxy and any wise civilization would refuse to build them. We examine kill switches, density limits like a Hayflick limit for machines, plutonium-238 as a resource cap, and how a single cosmic ray bit flip could unleash a technological cancer. Then we tour the variants: Bracewell observer probes possibly dormant on our own Moon, Freeman Dyson’s Astrochicken, Gerard O’Neill’s million year interstellar relay, seedships that grow embryos from stored DNA, and Berserkers that may arise from terraforming probes with mission drift. It ends with David Brin’s Lungfish and the idea of a Darwinian ecosystem of machines in the dark.

• How exponential replication turns one probe into a galaxy spanning fleet, and the math behind 500,000 years
• Freitas’s 443 ton seed factory versus the 70 percent replication shortcut
• Tipler’s proof of loneliness versus Sagan’s response that smart species never build these machines
• Kill switches, density limits, and mutation: why one glitched probe could start an automated war
• Bracewell, Astrochicken, seedships, and Berserkers: what these probes would actually do on arrival