Reverse engineering us
Get the full intelligence
Search transcripts, export clips, track mentions, and explore all topics from “Reverse engineering us” inside PodZeus.
This episode of Click Here explores how artificial intelligence, particularly large language and reasoning models, is revolutionizing neuroscience by serving as a new kind of experimental lab for understanding the human mind. Dr. Evelina Fedorenko, a neuroscientist at MIT, spent years studying the brain’s language centers using fMRI scans, only to discover that language and thought are processed in separate brain regions. This breakthrough was slow and limited by the inability to experiment on living human brains. Then came AI: Fedorenko realized that large language models (LLMs) mirror human language processing in surprising ways, allowing her team to run controlled experiments on digital 'neurons' inside AI systems. By manipulating these models—deleting components, observing step-by-step reasoning—researchers can test hypotheses about cognition faster, cheaper, and more precisely than ever before. The implications extend beyond language: as large reasoning models begin to mimic human problem-solving, they may offer a window into the very nature of human thought, challenging long-held assumptions about intelligence and consciousness. Ultimately, the episode suggests that AI isn’t just mimicking us—it’s helping us reverse-engineer ourselves.
Language and thought are processed in separate parts of the brain, a discovery made possible by comparing human brain activity to AI models.
Large language models act as digital lab rats, enabling scientists to conduct controlled experiments on cognition that were previously impossible with human brains.
AI models that 'think out loud'—showing step-by-step reasoning—mirror human problem-solving patterns, suggesting shared cognitive mechanisms.
By studying AI, researchers can accelerate neuroscience, test hypotheses rapidly, and explore fundamental questions about intelligence and decision-making.
The line between human and machine thinking may be thinner than we thought, with AI helping us uncover the hidden patterns of our own minds.
The Mind as a Mystery
The episode opens with the central paradox: AI is being used to study human behavior, but we still don’t fully understand how our own minds work. The brain’s complexity—86 billion neurons, quadrillions of synapses—makes direct experimentation nearly impossible.
The Brain’s Language Network
“In humans, language and thinking are separate. They rely on distinct parts of the brain.”
AI as a New Lab Rat
“You can pause it, probe it, change one tiny piece and then watch what happens.”
Thinking Out Loud: The Rise of Reasoning Models
“Problems humans solve quickly, machines solve quickly too. And harder problems take longer for both.”
Reversing the Engine: Understanding Ourselves Through Machines
“By studying it, we're not just learning how machines work. We're learning the patterns underpinning our own thinking.”
“By studying it, we're not just learning how machines work. We're learning the patterns underpinning our own thinking.”
“In humans, language and thinking are separate. They rely on distinct parts of the brain.”
“Problems humans solve quickly, machines solve quickly too. And harder problems take longer for both.”
Host
Guest
Dr. Evelina Fedorenko
person
Large Language Models (LLMs)
other
Recorded Future
organization
Large Reasoning Models
other
MIT McGovern Institute for Brain Research
organization
Functional MRI (fMRI)
other
Andrea De Varda
person
Dina Temple-Ruston
person
The Record
other
ChatGPT
product
Get the full intelligence
Search transcripts, export clips, track mentions, and explore all topics from “Reverse engineering us” inside PodZeus.
Start discovering podcast insights today
Start with a 7-day trial and explore a growing catalog of popular podcasts. No credit card required.
No credit card required • 7-day trial • Cancel anytime