The 1943 Paper That Sparked AI Just Got a Major Rewrite

Seventy years ago, two scientists named Warren McCulloch and Walter Pitts wrote a paper describing brain cells as logical switches that could compute. That 1943 work is often called the starting point of artificial intelligence. Now, a researcher named Nima Dehghani has written a book chapter that digs into that original paper—and shows the popular version of the story is a cartoon.

Dehghani explains that McCulloch and Pitts were doing more than saying a neuron is on or off. Their model included inputs that can completely block a signal, like a veto. They also studied networks with loops, which don't compute forever but settle into repeating patterns. And they looked at how a network could recognize the same thing even when it moved or changed—a problem still central to today's image recognition software. The chapter also revisits their experiments on frog retinas, which revealed that the eye itself does a lot of image processing before the brain even sees it.

Why should you care? Because this corrected history explains the foundations of the AI hiding inside your phone, search engine, and self-driving car prototypes. The frog retina work, for instance, is a direct ancestor of today's computer vision. The chapter also shows a clever lesson: if scientists only measure a neuron's average response, they may wrongly conclude it does nothing special. Method problems can hide real activity—a warning that applies to both brain science and AI development.

The chapter is part of an open book called 'NeuroAI,' and it includes code so readers can recreate the figures themselves. For the rest of us, it's a rare chance to see that AI's origin story is more subtle and more human than the simple binary-switch myth—and that a frog's eye still has secrets worth stealing.