The Road Not Taken: Walter Pitts, IBM, and the 3D Revolution
I recently read about IBM's new sub-1-nanometer chip technology, and it triggered a memory from Dark Hero of the Information Age. It’s strange when a modern engineering breakthrough illuminates a ghost from the 1940s.
The Ghost of Walter Pitts
After his famous 1943 paper modeling neurons as binary logic gates, Walter Pitts realized the brain wasn't just all-or-nothing switches. He began work on his doctoral dissertation, envisioning "a layered device, a three-dimensional net." He explored this using analog, continuous mathematics—a bold leap beyond flat, binary logic.
Norbert Wiener was captivated. But after a bitter falling out with him, Pitts simply stopped. He burned the manuscript. It became a "road not taken."
IBM’s "Nanostack"
Fast forward to 2026. IBM announces their sub-1nm chip. The breakthrough isn't just making things smaller; it's making them vertical. Their "nanostack" architecture uses 3D sequential integration to stack nanosheets. For decades, the industry fought to shrink transistors horizontally on a 2D plane. IBM's solution is to move into the third dimension.
The Convergence
I'm not an engineer, but conceptually, the parallel is striking. Both represent a paradigm shift born out of hitting a wall.
For Pitts, the wall was logical: a flat, 2D network of binary neurons couldn't capture the brain's analog complexity. He needed a 3D space for continuous, nuanced interactions.
For the semiconductor industry, the wall was physical: you can only shrink a 2D grid so far before quantum tunneling ruins the party. IBM realized they needed to build up rather than just out.
One was a theoretical biological vision; the other is a physical engineering feat. But both independently arrived at the same conclusion: when you hit the limits of the 2D world, the third dimension is the only way forward.
The Echo
We often think of progress as a straight line, but sometimes it's a spiral. We loop back to old ideas, rediscovering them in new contexts. Walter Pitts never got to see his 3D net come to life. But in IBM's nanostack, the ghost of his idea lives on.
If you found this useful, consider supporting my work.