Origins to Now traces the complete, 75-year story of artificial intelligence — from the first serious argument that a machine could think, to the systems now embedded in your everyday workflow.
The series runs in five acts (chapters). Act I asks whether "thinking" can even be defined. Act II covers the two winters that nearly ended the field twice. Act III covers the shift from hand-coded rules to machines that learn from data. Act IV covers the deep learning and transformer breakthroughs that made modern AI possible. Act V lands on the present — and the questions nobody's answered yet.
Twenty-one episodes total, one new episode every Monday. Below is the full index, updated as each episode goes live — follow along in order, or jump straight to whichever chapter you're curious about.
Can Machines Think? The Question That Started Everything (12:15)
Ep. 01 — Can Machines Think? The Question That Started Everything
In 1950, Alan Turing asked whether a machine could think — then sidestepped the question entirely, replacing it with a test instead: can it fool a human? That move became the founding logic of AI as a field. This episode covers where it came from, and why it matters more now that modern AI already passes his test. [Watch on YouTube →]
The Thinking Machines Before Computers: Leibniz, Boole, and the Logic of Thought (12:26)
Ep. 02 — The Thinking Machines Before Computers: Leibniz, Boole, and the Logic of Thought
Turing's question didn't come from nowhere — it was the endpoint of two centuries of mathematicians trying to turn thought into a formal system. This episode traces that line, from Leibniz's dream of a universal logic to Boole's algebra of thought, up to 1950. [Watch on YouTube →]
The Turing Test: How One Thought Experiment Defined a Field (15:41)
Ep. 03 — The Turing Test: How One Thought Experiment Defined a Field
Episode 1 introduced the question. This episode goes deeper into the test itself — how the Imitation Game actually works, why Turing built it the way he did, and why a thought experiment from 1950 still shapes how we argue about AI today. [Watch on YouTube →]
The Birth of AI: The 1956 Summer That Named a Revolution (17:12)
Ep. 04 — The Birth of AI: The 1956 Summer That Named a Revolution
In the summer of 1956, a small group of researchers gathered at Dartmouth College convinced that machine intelligence was a summer's work away. They were wrong about the timeline — but this is the meeting that gave the field its name, and its founding optimism. [Watch on YouTube →]
The Optimism Era: Early AI Programs and Why Everyone Thought We Were Almost There (17:50)
Ep. 05 — The Optimism Era: Early AI Programs and Why Everyone Thought We Were Almost There
The first AI programs could prove theorems and play checkers, and researchers assumed human-level intelligence was just a few years out. This episode covers what those early systems actually did — and why their success created expectations the field couldn't keep. [Watch on YouTube →]
The First AI Winter: When the Money Ran Out and the Hype Collapsed (18:13)
Ep. 06 — The First AI Winter: When the Money Ran Out and the Hype Collapsed
By the early 1970s, the promises hadn't materialized, and funding dried up almost overnight. This episode covers what actually broke down — the real technical limits behind the collapse, not just the popular myth of AI "failing." [Watch on YouTube →]
Expert Systems: The Second Wave and Why Rules Weren't Enough (18:59)
Ep. 07 — Expert Systems: The Second Wave and Why Rules Weren't Enough
AI came back in the 1980s with a new approach: encode an expert's knowledge as explicit rules. It worked well enough to build a real industry — until the same rigidity that made it work also made it collapse. This episode covers both halves. [Watch on YouTube →]
The Second AI Winter: What Went Wrong the Second Time (16:49)
Ep. 08 — The Second AI Winter: What Went Wrong the Second Time
Rule-based systems failed for different reasons than the first winter did — and the field's second collapse taught a harder lesson than the first one. This episode covers what actually changed heading into it, and why AI needed a completely different foundation to recover. [Watch on YouTube →]
Neural Networks: The Idea That Refused to Die (15:11)
Ep. 09 — Neural Networks: The Idea That Refused to Die
Neural networks were proposed decades before they worked, dismissed as a dead end, and quietly kept alive by a small group who refused to abandon them. This episode covers why the idea persisted — and what finally made it viable. [Watch on YouTube →]
Backpropagation: The Algorithm That Made Neural Networks Work (13:54)
Ep. 10 — Backpropagation: The Algorithm That Made Neural Networks Work
Neural networks needed one missing piece before they could actually learn: a way to correct their own mistakes. Backpropagation was that piece. This episode covers how it works, in plain English, and why it became the algorithm the entire deep learning era is built on. [Watch on YouTube →]
The ImageNet Moment: How a Dataset Changed Everything in 2012 (9:51)
Ep. 11 — The ImageNet Moment: How a Dataset Changed Everything in 2012
A neural network trained on a massive labeled dataset didn't just win an image recognition competition in 2012 — it beat every other approach by a margin nobody expected. This episode covers what ImageNet actually was, why the win mattered so much, and how it reset the direction of the entire field almost overnight. [Watch on YouTube →]
The Hardware Revolution: Why AI Needed GPUs Before It Could Think (11:12)
Ep. 12 — The Hardware Revolution: Why AI Needed GPUs Before It Could Think
Neural networks existed for decades before they became useful — the missing piece wasn't a better idea, it was enough raw computing power to run one at scale. This episode covers how a chip built for video games ended up powering the deep learning revolution, and why hardware, not just theory, decided when AI finally worked. [Watch on YouTube →]
Deep Learning: What It Is, Why It Worked, and Why Now (12:57)
Ep. 13 — Deep Learning: What It Is, Why It Worked, and Why Now
A network that teaches itself what to look for, one layer at a time. What "deep learning" actually means, and why 2012 didn't stop at one good result.
Word2Vec and Embeddings: When Words Became Vectors (12:14)
Ep. 14 — Word2Vec and Embeddings: When Words Became Vectors
How Word2Vec turned words into points in space, why "king minus man plus woman" lands on "queen," and the one flaw that sets up the next big idea in AI.
The Attention Mechanism: The Idea Behind the Transformer (13:57)
Ep. 15 — The Attention Mechanism: The Idea Behind the Transformer
How a network learned to let every word weigh every other word in a sentence and rebuild its own meaning on the fly, the actual mechanism behind the word "attention."