KI kommt (großteils) nicht aus dem Silicon Valley
Der Betriebswirt Andrew McAfee schreibt:
It’s AI, right? Surely, Silicon Valley’s most important innovation is AI.
No, for the simple reason that AI and machine learning aren't Silicon Valley innovations. The “Dartmouth Summer Research Project on Artificial Intelligence,” which has been called the “Constitutional Convention of AI,” was held in 1956 in New Hampshire. The proposal to the Rockefeller Foundation to secure funding for the conference (and seems to have coined the phrase “artificial intelligence”¹) was drafted by four East Coasters, and its list of people interested in the embryonic field contains no Northern California addresses or organizations.
Most of the major advances in the history of AI and ML were developed far from the Valley. As far as I can tell, in fact, the first big AI innovation that came out of a Valley company (as opposed to an organization like DeepMind that was acquired by a Valley company) was the Transformer architecture that underlies much of current Generative AI. This architecture was proposed in the landmark “Attention Is All You Need” paper, published in 2017 by a team of Googlers. (correct me if I’m wrong here, please; I’m not a historian of AI).
So Silicon Valley can’t claim AI. But (foreshadowing alert) it can correctly claim to be the home of most of the companies leading the current AI boom. And Web 2.0, the Cloud, mobile computing, the online economy, the chips that power the online economy, and most of the other important things that have happened this century at the intersection of capitalism and technological progress.
It’s remarkable how concentrated the US high-tech industries are. Virtually all the largest and most valuable tech companies in the US except for Microsoft and Amazon are crammed into a tiny patch of Northern California real estate. What’s going on there?
I favor the Markoff-Saxenian-Mallaby hypothesis, which I just formulated. In his book What the Dormouse Said veteran tech journalist John Markoff takes us back to NoCal at the dawn of the computer era and shows us the improbable mix of wooly academics, campus radicals, pocket protector-wearing defense sector engineers, and frequently-tripping hippies who came together because they were all interested in these fascinating new machines. The digital (counter)culture that arose is still very much part of the Valley, as epitomized by the number of excellent computer scientists who stagger around Burning Man every year.
Sociologist AnnaLee Saxenian highlighted the knowledge-sharing and -creation benefits provided by this freeflowing culture, contrasting it with the stable, hierarchical, buttoned-down culture of Boston-area tech firms. Believe it or not, young people, there was a time when it was unclear whether greater Boston or NoCal was going to be America’s high-tech Mecca. Saxenian argued convincingly that Silicon Valley won in large part because people moved around more easily there, and hence so too did important ideas and innovations. In The Power Law, financial writer Sebastian Mallaby stresses how important the Valley’s venture capitalists have been to all that cross-pollination, and how they too left behind their their fusty East Coast peers.
Once the Valley got a head of steam, agglomeration economies kicked in. These are “the benefits that come when firms and people locate near one another together in cities and industrial clusters,” as economist Ed Glaesser puts it. When the benefits are big enough the cluster can be durable; think about New York and London for finance, or Detroit for carmaking (foreshadowing alert #2). So Silicon Valley remained Silicon Valley because most of the people who knew how to do the Silicon Valley thing of building successful tech firms were in Silicon Valley.
Hier der ganze Beitrag:
