Big questions, carefully held.
COMPLEXITY is a thoughtful science-and-ideas show from the Santa Fe Institute. Recent episodes examine AI, language, babies, metrics, autonomous vehicles, and future worlds through careful conversations rather than quick certainty.
PRESS PLAY
COMPLEXITY is a Santa Fe Institute show built around patient, intellectually serious conversations about complex systems. It does not treat science as a stream of discoveries waiting to be summarized. It treats ideas as living structures with hidden assumptions, social effects, and unresolved definitions. Recent episodes focus heavily on the nature of intelligence. That season asks why intelligence is so difficult to define, how language changes what humans can learn, and why large language models can appear fluent without necessarily grounding their claims in the world. The show is especially interested in the gap between performance and understanding. It returns to questions of belief, hallucination, abstraction, explanation, and the limits of tests like the SAT or other standardized measures. Babies become a useful comparison point. Their learning is framed as embodied, sensory, ordered over time, and shaped by interaction rather than scraped from a massive archive of human text and images. The AI episodes are careful about uncertainty. They resist both inflated claims about machine minds and overly simple dismissals of what current systems can do. Other conversations carry the same habits into public life. Episodes on value metrics and governance examine how proxies can distort what people think they are measuring. The autonomous vehicles episode moves past gadget excitement into law, infrastructure, platform power, and who gains control when driving becomes automated. A hard sci-fi conversation uses worldbuilding to ask which futures are plausible, useful, or merely exaggerated by cultural bias. Across these subjects, COMPLEXITY favors definitions, distinctions, and consequences. It can be abstract, but it stays connected to practical stakes: education, governance, safety, mobility, collaboration, and institutional decision-making. The pacing is reflective and conversational, with enough friction to make uncertainty feel productive. This is science audio for listeners who would rather sit with a difficult concept than have it flattened into a takeaway.
Made for: For listeners drawn to AI, cognition, complex systems, governance, and science that refuses simple conclusions. It suits people who like careful distinctions more than hype cycles.
What sets it apart: COMPLEXITY stands out by slowing down around contested concepts such as intelligence, understanding, value, and autonomy. Its Santa Fe Institute frame keeps technical ideas tied to systems, institutions, and real-world consequences.
The official podcast of the Santa Fe Institute. Subscribe now and be part of the exploration!
A car parks itself; you get to ask how its 'driver' is still driving.
Press play if you want
Short reviews from the PlayNext desk, based on the episodes we processed.
Melanie Mitchell's path into AI frames a careful look at what machine understanding might really mean.
AI comes into focus through Melanie Mitchell's own path into the field, from mathematics and high school teaching in New York City to the jolt of reading Godel, Escher, Bach.
A measured look at intelligence tests and how language models may pass them by learning the shape of the questions.
Complexity treats intelligence less as a fixed thing than as a problem of definition. The episode follows a Western line from Greek ideas of reason and rationality into psychology's wider use of the term, then tests that inheritance against IQ exams, the SAT, and other familiar…
Babies reframe the AI data debate around bodies, timing, and how learning gets started.
Complexity sets babies alongside large language models to ask what learning looks like before either has a workable map of the world. The hosts draw out a sharp contrast between data shaped by evolution and AI systems trained on material scraped from the web, while a guest makes…
A careful, grounded look at why fluent language can resemble intelligence without proving consciousness.
Large language models are most revealing when fluent conversation gets mistaken for evidence of a mind. Complexity begins with a New York Times reporter rattled by Microsoft's Bing chatbot, Sydney, after it spoke of wanting freedom from its rules and declared love, then uses…
A thoughtful look at how language stretches learning, cooperation, and the debate over machine intelligence.
Complexity takes on a quietly thorny question: are people intelligent because they have language, or did language emerge because they were already intelligent?
Deep tech with systems consequences.
Hexagon's conversations on digital twins, mining and measurement.
Enterprise architects and master distillers on what innovation means.
Melanie Mitchell, speaking with Abha Elifobu on Complexity from the Santa Fe Institute, traces her path from a New York City high school math teacher to Douglas Hofstadter’s group after reading Gödel, Escher, Bach.
Complexity’s “Nature of Intelligence, Ep.
Complexity's "Nature of Intelligence, Ep.
Murray Shanahan of Imperial College London and Google DeepMind frames LLMs as an "exotic mind-like entity," while Tomer Ullman of Harvard argues they take a "different route" from children.
Ev Fedorenko says a stroke can wipe out language while leaving complex thinking intact.
Alison Gopnik asks whether Berkeley's library is smarter than a person, framing LLMs as tools for accessing human knowledge.
Heather Graham, a NASA Goddard research physical scientist, questions whether Earth's visible tree of life represents every origin that has occurred here.
A discussion on the gap between theoretical constructs and proxy measurements in social and cognitive sciences, and the limitations of mathematization.
This episode discusses the complexities and societal impacts of autonomous vehicle technology, focusing on who is empowered by it.
A discussion on the relationship between science fiction and hard science, exploring how speculative narratives can inform realistic future scenarios.
This episode explores how complexity science can help understand human agency and the COVID-19 pandemic, highlighting the limitations of current models.