How the World Got Wired · A Naval Office in Arlington

The Pattern, Again

You should not be surprised by this. You have seen this pattern many times in this book.

The Air Force paid for Paul Baran to invent packet switching. ARPA paid for Licklider to imagine the intergalactic computer network. ARPA paid for BBN to build the IMP and the original ARPANET. The National Science Foundation paid for the public computer labs where Marc Andreessen wrote Mosaic. CERN, an international physics laboratory funded by European governments, paid for Tim Berners-Lee to invent the web.

Now the Navy paid for the backpropagation paper.

Every single foundational piece of technology in this book was, in some way, paid for by a government — by taxpayers, by the public, by people who had no idea what their money was being used for. Each piece was then released, mostly for free, into the world, where private companies eventually built enormously valuable businesses on top of it.

This is not, by itself, a moral statement. It is just a description. The pattern is there. You should notice it.

There is one more thing I want to tell you before we leave this chapter, because it points to the next one.

The Office of Naval Research, when it funded the PDP group’s work in the 1980s, was acting on a long-standing assumption: that basic research in cognitive science and artificial intelligence might, eventually, help the Navy. The Navy did not know how. The Navy was not trying to build chatbots. But the Navy was paying, partly, because eventually the work might lead to systems the Navy could use — for surveillance, for decision-making, for tactical analysis, for weapons.

Not all the scientists who took the Navy’s money were entirely comfortable with this.

The young British scientist named Geoffrey Hinton, who had been at the center of the PDP group, who had been one of the three authors of the backpropagation paper, who would eventually be called the godfather of deep learning — Hinton, for reasons that had partly to do with his family history and partly to do with his own values, had been thinking about this for some time.

In 1987, the year after the Nature paper came out, he made a decision.

He left the United States. He took a job in Canada.

He went, partly, because he did not want his work to be paid for, anymore, by an American military office.

That’s the next chapter.

All 24 chapters of How the World Got Wired · Library · All courses