Listen

Description

In this eleven-minute Episode 33, Brick Thompson and Ed Daniels discuss Google’s new AlphaEvolve, an evolutionary coding agent developed by Google DeepMind. AlphaEvolve uses an LLM to generate different versions of the target algorithm’s code, and then applies an evolutionary algorithm to evaluate and select the most promising code version.

Ed and Brick riff on how this tool could be used in the business world to improve business processes. They also speculate on how humans fit into this approach. Do we still have a role to play? Brick and Ed wrap it up by spinning off into a discussion of the Singularity and how it sometimes feels like we are in the middle of it right now.

·       See more on AlphaEvolve at https://deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/ 

·       If you really want to dive into the topic of AI self improvement, read this article by Carlos Perez https://medium.com/intuitionmachine/the-space-of-transcendental-machines-from-population-dynamics-to-recursive-self-improvement-5a3a047c1534

The name of this podcast, Beyond Electric Sheep, is a reference to the novel Do Androids Dream of Electric Sheep? by Philip K. Dick. That novel is the source for the movies Blade Runner (1982) and Blade Runner 2049 (2017). It explores the themes of artificial intelligence, empathy, and reality in a dystopian future where humans and androids coexist. 

The novel raises the question of whether androids can dream or feel anything at all, and whether they are different from humans in any meaningful way. In our Beyond Electric Sheep podcast we go beyond the novel and discuss the broader applications of artificial intelligence in our personal lives and in our businesses. We explore the pragmatic as well as the ethical and philosophical issues that arise from creating and interacting with intelligent machines.

Beyond Electric Sheep is also available on YouTube at https://www.youtube.com/playlist?list=PLYg2J0jET_BPCKSnsxz8mleEmYogYJZhn