(www.plasticpals.com) The Machine That Changed The World (aka The Dream Machine in the UK) isan exhaustive 5 part documentary about the history of computing produced by WGBH Television and the BBC in 1992. It was saved from the dustbins of history by Andy Baio (waxy.org) who digitized it and made it available online. The fourth episode, The Thinking Machine, focuses on artificial intelligence and begins its exploration in the late 1950s when Marvin Minsky and John McCarthy set up the A.I. department at MIT.
In the early 1960s they were so convinced they were on the right path that many believed thinking computers and robots not unlike those in science fiction were only 5 to 10 years away. Of course, when they tried to get their first simple robots to interact with the real world everything fell apart, and early chatbots and machine translators were woefully inept. Tasks which are effortless for the human brain proved exceedingly difficult to replicate. Humorous anecdotes are accompanied by fascinating clips of past projects. When Hubert Dreyfus published his book What Computers Can't Do: A Critique of Artificial Reason, some predicted that artificial intelligence was doomed.
Yet despite the failures, there were promising avenues of research. Some moved on to narrow artificial intelligence, which proved useful in very specific domains. Others tried to catalogue and categorize knowledge in an attempt to reproduce common sense. Others believed that embodiment was the only way for a computer to possess real knowledge. And there are those who have believed from the beginning that the best way to make a thinking computer is to simulate the brain. Early neural networks were called Perceptrons (now there's a term I'd like to see used more often!).
As the documentary catches up to the late '80s, things start to get familiar. Autonomous cars, text-to-speech synthesizers, and even long-running projects like Cyc begin to appear.
In the early 1960s they were so convinced they were on the right path that many believed thinking computers and robots not unlike those in science fiction were only 5 to 10 years away. Of course, when they tried to get their first simple robots to interact with the real world everything fell apart, and early chatbots and machine translators were woefully inept. Tasks which are effortless for the human brain proved exceedingly difficult to replicate. Humorous anecdotes are accompanied by fascinating clips of past projects. When Hubert Dreyfus published his book What Computers Can't Do: A Critique of Artificial Reason, some predicted that artificial intelligence was doomed.
Yet despite the failures, there were promising avenues of research. Some moved on to narrow artificial intelligence, which proved useful in very specific domains. Others tried to catalogue and categorize knowledge in an attempt to reproduce common sense. Others believed that embodiment was the only way for a computer to possess real knowledge. And there are those who have believed from the beginning that the best way to make a thinking computer is to simulate the brain. Early neural networks were called Perceptrons (now there's a term I'd like to see used more often!).
As the documentary catches up to the late '80s, things start to get familiar. Autonomous cars, text-to-speech synthesizers, and even long-running projects like Cyc begin to appear.