AI History · EP 01

AI died twice.
And it was revived twice.

In July 1958, the New York Times reported on the unveiling of a "machine that learns by doing." Eleven years later, a single book delivered its death sentence. Thirty years after that, a single Nature paper brought it back.

5 min read 2026.05.04 1958 → 1986

01July 1958, the New York Times

July 8, 1958. The New York Times ran a prominent story on a US Navy demonstration in Washington — "NEW NAVY DEVICE LEARNS BY DOING." The press release contained this striking claim:

"The Navy revealed the embryo of an electronic computer today that it expects will be able to walk, talk, see, write, reproduce itself and be conscious of its existence."

— New York Times, 1958.07.08 (paraphrasing the Navy press conference)

The man behind the announcement was Frank Rosenblatt, a 30-year-old psychologist at the Cornell Aeronautical Laboratory. The machine he built was a roughly cabinet-sized device with about 400 photocells, called the Mark I Perceptron — the first hardware implementation of an artificial neural network.

🧠
Frank Rosenblatt
1928–1971 · Cornell Aeronautical Laboratory · Psychologist + computer scientist

The first person to take the hypothesis "a machine can recognize patterns like a human can" and build it in hardware. The Mark I could be shown letters and learn to classify them as 'A' or 'B'. The US Office of Naval Research funded the work.

The perceptron's operation is simple. Multiply each input by a weight, sum them, and if the sum crosses a threshold output 1, otherwise 0. If the prediction is wrong, nudge the weights a little. Try again. Adjust again. That's it. But the moment that loop ran, "a machine learning from data" had become a working concept for the first time.

021969, one book ended everything

Eleven years later, in 1969, two giants at MIT — Marvin Minsky and Seymour Papert — published a book simply titled Perceptrons. The cover featured two patterns and asked a simple question: "Are these the same?"

📖
Marvin Minsky & Seymour Papert
MIT · "Perceptrons" (1969)

The book proved mathematically that a single-layer perceptron could not solve nonlinear problems like XOR. That is, it couldn't even learn the simple logic of "true when exactly one input is true."

The fix, in principle, was clear: stack multiple layers. But that opened a harder question — "how do you train all the weights in a multi-layer network simultaneously?" Nobody had an answer.

⚠️ The first AI Winter begins
Funding dried up. The US government, the UK government, and academic departments all turned away from neural networks. The next 15 years were essentially dead time for the field. Frank Rosenblatt died in a boating accident on Chesapeake Bay in 1971 — on his 43rd birthday.

031986, six pages in Nature

October 1986. Nature, volume 323, pages 533–536. A six-page paper titled "Learning representations by back-propagating errors." Three authors.

David Rumelhart · Geoffrey Hinton · Ronald Williams
UC San Diego + Carnegie Mellon · Nature 1986

The core idea, in one sentence — "apply the chain rule of differentiation in reverse, from output back to input, to compute gradients for every weight in one pass." This is what we now call backpropagation.

The implication was immediate. Multi-layer networks could now be trained. Minsky's 17-year-old objection — that perceptrons can't learn XOR — was resolved. The door to deep neural networks was finally open.

"The algorithm is so robust that 40 years later, every neural network still trains the same way."

— ChatGPT, GPT-4, Stable Diffusion, Claude — all trained with backprop

04And then it slept again

Backpropagation didn't trigger an immediate AI explosion. Throughout the 1990s, neural networks stayed at the margins. Two reasons:

Problem ①
Not enough data
Training a serious neural network demanded hundreds of thousands of labeled examples. In the 1990s digital photographs themselves were rare.
Problem ②
Computers were too slow
A single training run took weeks on CPUs. SVMs, decision trees, and HMMs — all simpler — outperformed neural networks in industry.

So neural networks went through a second AI winter from the late 1990s through ~2010. Only a handful of researchers — Hinton, Yann LeCun, Yoshua Bengio — kept the embers alive. They would later be called the "godfathers of deep learning" and share the 2018 Turing Award.

🔥 Then in 2012, everything changed
Two of Hinton's students entered the ImageNet competition. Their names were Alex Krizhevsky and Ilya Sutskever. Their model — just called 'AlexNet' — cut the error rate from roughly 26% to 16% in a single shot. From that day on, AI never went back to sleep.

05So when we say AI died twice

Once in 1958, again in the 1990s. And each time, one person and one paper brought it back. In 1986 it was Hinton's backpropagation. In 2012 it was Krizhevsky's AlexNet.

The ChatGPT you use today, Stable Diffusion, autonomous driving, semiconductor fab AI — all of it traces back to Rosenblatt's cabinet-sized machine in 1958. And the way they all learn is exactly what Hinton wrote down in 1986.

In the next post (EP02), we step into the era that started in 1989 when Yann LeCun got a network to read handwritten ZIP codes at Bell Labs. How did his "LeNet" eventually become the camera in your phone, 30 years later?

🧪
Try it · AI Lab
Run one cycle of neural network training yourself →
Step through Forward · Backward · Update on a 2→3→1 net and watch the loss decrease in real time.
AI History · Series Navigation
← Previous
First episode of the series
Next →
EP02 · How Computers Got Eyes (coming soon)