A Human Algorithm With Flynn Coleman

December 6, 2019·26 min·2 clips

Why Raytheon is studying octopus intelligence to build better robots than human-brain AI.

Brian is still trying to get his arms around AI, and that makes the interview work. He is informed, but he does not act like the whole thing is obvious. Coleman brings the conversation back to human rights and technology, with a premise that is broad without floating away: what does it mean to be human when powerful tools can be used for the common good, or for something less generous? The language gets messy quickly. Coleman points out that even people inside the field do not always agree on basic definitions, including intelligence itself. Neural networks are the neat example that turns out not to be so neat. The usual explanation reaches for the human brain, but Coleman is careful about that analogy. It helps people picture the idea, but it is not exact. She talks about mimicking the vertebrate brain cortex, which is more precise and less magical. That is where the episode gets useful. If the terms are shaky, the public confusion around AI starts to feel pretty reasonable. Coleman says artificial intelligence may eventually fall away as a label, and notes that some people prefer augmented intelligence. That is not just wordplay. The name changes how people imagine the technology, who controls it, and whether humans are still in the picture. Brian keeps the conversation analytical rather than breathless. He treats A Human Algorithm as a grounding text on AI history, current capabilities, and ethical or social frameworks, not as a futurist sermon. Coleman keeps returning to responsibility. For her, AI is a social design problem as much as an engineering problem, with human rights and the common good close to the center. The global section is direct. China and the United States come up as rivals in what Coleman calls a real arms race, but she also notes examples of cooperation. Decolonization enters too. Her point is that the people building these systems should include more diversity, inclusion, representation, and decolonization. The ending returns to the book, with Brian calling A Human Algorithm the best AI grounding text he has read. It sounds less like a blurb than a recommendation with a job to do.

As heard by us

A thoughtful AI conversation that treats the technology as a human rights and public responsibility issue.

Brian McCullough uses this Techmeme Ride Home bonus episode to ask why AI deserves more than gadget coverage, and Flynn Coleman answers from a human rights frame.

Read the full review in PlayNext →

Why you'd press play

For a clearer AI frame, let Flynn Coleman reset the terms around what intelligence means.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Techmeme Ride Home