Explore
A curated path through the most important questions about AI and what it means for us. Videos, books, podcasts, and essays — following a natural curiosity arc.
# What is Intelligence?
Before asking whether AI is intelligent, challenge the assumption of what intelligence even means.
Michael Levin · Lex Fridman Podcast · 2022 Video
Intelligence exists far beyond brains — in cells, embryos, and biological collectives that solve problems without any nervous system. A mind-expanding conversation about what intelligence really means.
Joscha Bach · Lex Fridman Podcast · 2023 Podcast
Cognitive scientist and AI researcher Joscha Bach takes a sweeping journey through computation, consciousness, and what it means to have a mind.
Blaise Agüera y Arcas · Long Now Talks · 2024 Video
A deep exploration of intelligence beyond the AI lens — spanning biological, collective, and machine intelligence, and why the question matters more than the answer.
LLM Bento LLM Bento
Complex behavior from simple rules. Watch intelligence-like patterns emerge from three basic laws — in your terminal.
# Can AI Actually Understand?
The first thing anyone wonders after a surprisingly good ChatGPT response. Even the builders disagree.
LLM Bento LLM Bento
Tokens, attention, context windows — the building blocks of how language models actually process text.
Andrej Karpathy · 2023 Video
One of the best AI educators alive gives a clear, one-hour walkthrough of what large language models actually are, how they're trained, and where they're headed.
Grant Sanderson · 3Blue1Brown · 2017–2024 Video
Beautiful animated explanations of how neural networks learn, from basic building blocks to the transformer architecture behind modern AI. No prior background required.
Stephen Wolfram · 2023 Article
Step by step, the mechanics of how ChatGPT generates text and why it works as well as it does. A thorough, readable deep-dive freely available online.
Geoffrey Hinton · Romanes Lecture, Oxford · 2024 Video
The "Godfather of AI" explains why neural networks may genuinely learn deep representations of the world — and why digital intelligence could eventually surpass us.
Yann LeCun · Lex Fridman Podcast · 2024 Podcast
Meta's Chief AI Scientist explains why he believes current LLMs fundamentally cannot understand the world and why entirely new architectures are needed. A vital counterpoint to the hype.
Gary Marcus · 2022 Article
Deep learning has fundamental limitations — it cannot truly reason, understand causality, or reliably tell fact from fiction. A clear-eyed critique that has aged well.
Welch Labs · 2025 Video
A visual deep-dive into the inner workings of a small neural network we can fully interpret — making the abstract machinery of AI tangible and concrete.
LLM Bento LLM Bento
Does symbol manipulation equal understanding? The classic argument against machine comprehension, explored interactively.
# What Makes Human Thinking Different?
Once machine understanding gets complicated, you start questioning what understanding even means for us.
Iain McGilchrist · RSA Animate — 12 min · 2011 Video
How the two hemispheres of the brain offer fundamentally different ways of engaging with the world — and how civilization may be dangerously over-relying on one of them.
John Vervaeke · 50-episode lecture series · 2019 Video
How humanity lost its sense of meaning — drawing on cognitive science, philosophy, and contemplative traditions — and how we might recover it. An intellectual odyssey unlike anything else on YouTube.
The mind does not stop at the skull — our bodies and environments are part of how we think. A foundational book on embodied cognition.
Karl Friston on how the brain constructs its model of reality through predictive processing — and what consciousness and neuroscience reveal about the nature of thinking.
LLM Bento LLM Bento
Vervaeke's framework for human understanding — propositional, procedural, perspectival, and participatory knowing. Most of what makes us human lives beyond facts.
# Could AI Become Conscious?
The question that won't go away — and why it matters even if the answer is "no."
David Chalmers · TED Talk · 2014 Video
The philosopher who coined "the hard problem of consciousness" explains why subjective experience remains the deepest mystery in science. 19 minutes that reframe everything.
Could consciousness be more widespread than we think? Could it be an illusion? A short, accessible entry point into one of the biggest questions in science.
From plants to psychedelics to AI, Pollan traces the unmapped continent of consciousness — and argues that genuine thought requires embodiment and feeling, qualities machines fundamentally lack.
# What Are the Real Risks?
Now that you appreciate the complexity, the risks feel less sci-fi and more grounded.
Geoffrey Hinton · 2024 Video
In his Nobel Prize lecture, Hinton explains the ideas behind modern AI and closes with an urgent warning: we may be creating beings more intelligent than ourselves, with no idea how to stay in control.
Stuart Russell · 2019 Book
The standard approach to building AI is fundamentally flawed. Russell proposes machines designed to be uncertain about human preferences — making them inherently safer.
What happens when AI surpasses human intelligence? Scenarios from utopia to extinction. One of the most balanced introductions to AI's long-term implications.
Yuval Noah Harari · Frontiers Forum Keynote · 2023 Video
AI's real threat is not physical but linguistic — it has mastered the operating system of human civilization. A powerful talk from one of our most influential public thinkers.
80,000 Hours · 2025 Article
A carefully curated collection of the most influential writings shaping the debate on AI risk and safety. An ideal starting point.
Google DeepMind · 2025 Report
How the world's leading AI lab is planning for the technology it is building — covering misuse, misalignment, accidents, and structural risks.
# What Does the Future Look Like?
The optimistic vision and the sober reality check, side by side.
Dario Amodei · 2024 Article
What the world could look like if powerful AI goes right — breakthroughs in biology, medicine, economic development, and global peace. A detailed, hopeful counterweight to the doom narratives.
Dario Amodei · 2025 Article
The arrival of powerful AI as a turbulent rite of passage for civilization — five categories of existential risk and a concrete battle plan to navigate them.
Daniel Kokotajlo et al. · 2025 Report
A detailed forecast tracing a plausible path from today's AI to superintelligence, exploring both catastrophic and hopeful endings. Read by over a million people within weeks.
Leopold Aschenbrenner · 2024 Article
AGI by 2027, superintelligence shortly after. A former OpenAI researcher argues most people are not taking the speed of progress seriously enough.
Stanford HAI · 2025 Report
The most comprehensive annual survey of AI — research, industry, policy, public opinion, and societal impact. The reference for where AI actually stands today.
Nathan Benaich & Ian Hogarth · 2024 Report
An annually updated, data-driven overview of AI research, industry, talent, and geopolitics. More opinionated than Stanford's, freely available online.
More on LLM Bento
Tip: Ask the to recommend resources on specific topics.
