🥈 4/5 - The Chip: How Two Americans Invented the Microchip and Launched a Revolution

Read in Dec 2019
Book by T.R. Reid published in 2001

I try my best to write a short summary/review of the books I read, and this is one of them. I typically publish them on Goodreads, but also sync them to here.

Okay, Shackley & Co gave us the magnificent transistor in 1947, but how did we get from there to the general-purpose, spreadsheet-wrangling CPUs we have today? You can think of a transistor as a hose with an electric clamp. The clamp prevents water from flowing through it only when electricity goes to the clamp. If you stop sending electricity to the clamp, the flow of water resumes.

With the electric-clamp, we can now build circuits. Imagine a hose with two clamps next to each other, A and B:


If we send current through A, but not B, then put water through the hose, then
we won't get any water out:


But if we put water through A _and_ B, we'll get water out the other side:


That's called a "AND"-gate: a signal only passes through if both the 'clamps' (transistors) have power. We can construct other simple gates like OR-gates and NAND-gates this way. One of the modern heroes of computing from Bell Labs, Claude Shannon, wrote a thesis showing how we can build general-purpose machines by turning any problem into a combination of boolean functions. That means, we can solve any problem by putting enough transistors together, because we can form those gates with transistors. It means you can multiply numbers, add them together, so on, and so forth with transistors.

This worked well to create computers for a while, but as transistors shrunk and we continued to want them to get faster, a bottleneck crept in: wiring all those transistors together without error became near-impossible past a certain number of transistors. That threshold was reached in the 1960s. The problem was called "The Tyranny of Numbers." The circuit-engineers couldn't fathom how we'd put more transistors into our machines without error.

However, around this time, Jacky Kilby (Texas Instruments) and Robert Noyce (Fairchild Semiconductors) both independently came up with the idea of the integrated circuit: instead of wiring individual transistors together, put them on a silicon-oxide conductive surface. This meant no more tiny wiring. It meant much faster transistors, because they could get even smaller. The Tyranny of Numbers was a solved problem. Of course, since two people came up with it around the same time, legal hell ensued. That part is less interesting.

Robert Noyce, Goordon Moore, Andy Grove, and others left Fairchild Semiconducters where they had worked on integrated circuits and co-founded Intel. There, they made another magnificent leap into the world of microprocessors. Before this time, each customer wanted their own custom-designed chip. But now chips started becoming available that were general-purpose, like the Intel 8080.

And there you have it, stage for the Computer Age set!

Lovely books, especially in the heels of the one on Bell Labs.