Today, Vernor Vinge, Marvin Minsky and the future of computation. The University of Houston presents this series about the machines that make our civilization run, and the people whose ingenuity created them.
______________________
I was saddened to learn about the passing of Vernor Vinge. Dr. Vinge was a professor of computer science and mathematics, but was famous as a science fiction writer whose novels brimmed with new ideas that had a profound impact.
Vernor Vinge, 2006. Wikipedia
For example, Vinge’s 1981 novella True Names is a founding work of the cyberpunk genre. In 1981 home computers were a novelty, and the internet consisted of a couple hundred host machines accessible to a select few. Yet, in his novel Vinge envisioned a fully immersive, globally accessible cyberspace, computer hackers that bring down governments, and runaway artificial intelligence. He even suggested that our digital facsimiles can outlive us after we die. While these concepts are now familiar, it is hard to imagine what a reader in the 1980s would have thought of them.
Central to the book is the idea of an interface that hides the complicated workings of the machine. In Vinge’s book users enter an immersive medieval world to manipulate data. Today, we can access data and control our homes and cars through interfaces on our phones without knowing how to code.
The afterword to the book was written by Marvin Minsky, at the time the director of the Artificial Intelligence Laboratory at MIT. Minsky pointed out that the workings of our own mind remain mysterious. It is as if we use an interface to communicate our wants to our bodies: We speak and move our limbs without being aware of the intricate coordination of muscles that make speech and movement possible.
Marvin Minsky, 2008. Wikipedia
But automation may have dire consequences. Minsky suggested that tasks requiring intellectual effort, including programming, will soon be done by machines. We will voice our desires, and intelligent systems will translate them into reality. Minsky warns that such automation can be dangerous, as we rarely know exactly what we want, and we are not good at articulating our wants precisely. How can we make sure that machines understand our desires, so that we don’t end up like the sorcerer’s apprentice?
But Minsky’s biggest fear were machines that can improve themselves. He writes: “The ultimate risk … comes when we greedy, lazy master-minds are able at last to take the final step: to design … programs which are programmed to make themselves grow increasingly powerful”.
Minsky and Vinge wrote in the early 80s when supercomputers were thousands of times slower than today's smartphones, and skilled chess players could easily defeat the best programs. Yet, the era of intelligent machines was on the horizon. Now that era has arrived, and their warnings echo with striking accuracy.
This is Krešo Josić at the University of Houston where we are interested in the way inventive minds work.
(Theme Music)
_________________________________________________________________________________________________
The idea that artificial intelligence could slip out of human control was not new in 1982. For example, the 1966 novel Colossus by D.F. Jones envisioned a world dominated by super-computers: https://en.wikipedia.org/wiki/Colossus_(novel) For a list of fictional computers see here: https://en.wikipedia.org/wiki/List_of_fictional_computers
Comparing the speed of super-computers in the 1980s to modern smart phones: https://blog.adobe.com/en/publish/2022/11/08/fast-forward-comparing-1980s-supercomputer-to-modern-smartphone
This episode was first aired on June 25, 2024