The Dream Machine is an excellent and highly accessible introduction to the history of computing. It has excellent introductions of the big names - Von Neumann, Shannon, Englebart, etc - but avoids the temptation of "great man" history. A real standout is Chapter 2, which shows both how the "computer" originated from dozens of separate contributions like binary logic and vacuum tubes and how important integrators like Von Neumann were in piecing them together.
Reading The Dream Machine really gives one a sense for large a role the US military played in computing. War projects like ENIAC and Sage practically originated the programming profession, and DoD funding birthed desktop computing and the Internet almost as an afterthought. How much longer would computing or the Internet have taken to develop without WWII and the Cold War? Could private players like Bell Labs or academia have picked up the slack? How do Chinese subsidies to AI and hypersonics compare, and why did Japan fail to emulate these basic research subsidies in the 80s and 90s?
Sadly Licklider's ARPA and Xerox PARC seem impossible to create today. The corporate research lab is dead, and DoD funding just isn't how it used to be. And the loss isn't just financial:
- Contact with real-world problems spurs unforeseen innovation, like Shockley's invention of the transistor. While the Cambrian explosion of New Science funding is encouraging, does any private philanthropist have the sheer capital required to actually push bleeding-edge projects like SAGE into production?
- Bureaucratization and the broken grant / tenure process would make Licklider's hiring practices at ARPA illegal and grants to no-hopers like Englebart impossible More troublingly, fields like CS and physics may contain so much content that no human can piece abstractions like the computer together from disparate discoveries.
More troublingly, fields like CS and physics may contain so much content that no human can piece abstractions like the computer together from disparate discoveries.
The Dream Machine isn't perfect. Its narrative frame around JCR Licklider frankly feels contrived. Particularly in WWII and SAGE, Waldrop stretches a bit hard to fit Licklider's psychological research (though interesting) with the rest of the narrative. But at the end of the day, it's popular narrative history and shouldn't be judged for the limitations of the genre. It's not a policy prescription (see Ben Southwood / PARPA for that) and doesn't claim to be exhaustive. However, it's an excellent guide and orientation to the intellectual landscape of the postwar era and the field. Very much worth reading, especially if you know nothing about the field.