Escaping the Zelenograd Trap
On the parallels between the semiconductors wars of the 1900s, and the AI wars of today.
Chapter 8 of Chip War, Chris Miller’s popular book on the history of semiconductors, is aptly titled “Copy It”. This is a story that takes us back to Zelenograd, and reminds us why history never repeats itself, but it does often rhyme [1].
The Copy It Strategy
In 1958, the Soviet Union established Zelenograd, a “planned city for the development of electronics”. The impetus for setting up Zelenograd came from the US. Almost a decade earlier, the US’ industrialisation of the transistor had spurred a burgeoning electronics industry. The Soviets had taken notice of this success, understood the implications on national defence, and wanted to do something about it.
A few months after Zelenograd was commissioned, a Soviet spy studying in the US returned to Russia with an SN-51, one of the first integrated circuits sold in the US. The Soviet administration immediately ordered their engineers to “copy it, one-for-one, without any deviations”.
Soviet engineers reacted poorly to the suggestion that they should simply copy foreign devices. However, there was not much they could do as Shokin, the Soviet bureaucrat in charge of microelectronics, firmly believed in the “copy it” strategy.
The issue with Shokin’s strategy was Moore’s Law. Even if the Soviets managed to copy a design and acquire the necessary materials and machinery to produce it at scale, they would be behind. Moore’s Law was advancing so rapidly that copying designs was a hopeless strategy.
The Zelenograd Trap & AI
The Soviet’s “copy it” strategy condemned them to backwardness. As Chris Miller said, “Zelenograd was like Silicon Valley without the sunshine”. Whilst founders in Silicon Valley were building real stuff that people used, Soviet engineers were forced to copy old technology.
This brings me to our current moment in AI. Many people see a shiny piece of new technology and want to do something with it. There are countless “AI chatbots for [insert industry]” or “AI summarizers” on the market. I can’t help but think that many of these people are falling into the Zelenograd trap.
Copying can be a good strategy. But when the rate of innovation is so high (as we’ve seen with chip development, and now AI), a “copy it” strategy is likely to lead to consistently being behind the curve.
TSMC
Whilst the Soviets were preoccupied with old tech, the Taiwanese government (in particular Morris Chang) noticed something novel. He realised the world was ravenous for chips, and that we would need somewhere to produce them. So he set up TSMC – the first-ever fabless producer of chips. This meant that companies could design chips in-house, but outsource their production to TSMC.
In order to keep up with their demands, he had to keep pace with Moore’s Law. The result of his creation was the birth of one of the most valuable technology companies in Asia. It also made Taiwan the centre of a hotly debated conflict between the US and China.
What’s the takeaway?
The most consensus view in tech right now is: “Most of the value created by AI will be captured by incumbents.”
When Sputnik 1 launched, people assumed the Soviets had won the technology race. Then, American entrepreneurs built and scaled chip companies. After that, everyone assumed the US was in charge.
Again, they were displaced by a Taiwanese upstart. Today, TSMC produces 37% of our global chip supply.
There is a repeated cycle in history proving that great people at the edge of their fields create great companies. Every time there is innovation, there are new companies that grow to become very valuable. I don’t believe that the R&D departments at Google or Facebook control the future of AI. They may have distribution and resources, but they don’t necessarily hold the key to the future.
This is why I’m bullish on entrepreneurs building in AI. But there’s a valuable lesson to be learned from the Soviets. History suggests that you shouldn’t start with the shiny new technology. You need to start with a problem and find novel applications in areas where you have an advantage.
[1] Thanks for the great quote Mark Twain, although I tend to prefer this one by Ken Burns: “History doesn't repeat itself, but human nature remains the same.”
[2] For the growth stage founders reading this, the biggest risk to your company is a bad culture. Having a bureaucratic and secretive process that hinders the sharing of knowledge is a weakness. The 2-man teams moving quickly without any of this can (and will) achieve great things if you aren’t front-footed.