Thank you, all, for your engagement in our seminar and your participation on the first day of class. I’m very excited to be back in the classroom and look forward to our Mondays together. We are going to have so much fun this semester.
As I mentioned last Wednesday, we don’t expect you to carefully work through the optional readings. They’re there to add color and context to the main readings (and videos and other required stuff). This past week, for example, I couldn’t resist including Specifications for the Interconnection of a Host and an IMP. If you do nothing else, flip through the pictures at the start of the document and then visit the Harvard Mark I, located in the Science Center. Ok, well, visit it when we uncover it again — it’s currently being protected from the construction going on nearby. Sorry about that.
Flipping to the other optional reading, we talked about J.C.R. Licklider as a visionary and his importance to the development of the ARPANET. But beyond the ARPANET, Licklider was a strong proponent of cooperation between man and machine to produce a better world than either could alone. We will visit this topic again in Week 6, and for now, I just want to highlight the section in his paper that starts on page 10 under the subtitle, “Automatic Speech Production and Recognition.”
I’ve been underwhelmed in my interactions with Siri, but I’m still a techie at heart and can’t wait to get an Amazon Echo. It’s with this context that I read the first sentence in Licklider’s section on automatic speech recognition, which starts by asking “[h]ow desirable and how feasible is speech communication between human operators and computing machines?” In our time, quite desirable, right? We are just starting to scratch the surface of the potential uses of speech recognition, and many articles exist today talking about how Alexa is leading us into the future of intelligent homes.
Then I read the rest of Licklider’s section. He was far ahead of his time in focusing on speech recognition, but his view of the future was clearly clouded by the cultural norms and traditional business practices of his time. For instance, he argued the need for automatic speech recognition because “one can hardly take a military commander or a corporation president away from his work to teach him to type.” Notice the gender bias in this statement. Even if Licklider himself envisioned leadership roles for women, he later writes “that a corporation president would be interested in a symbiotic association with a computer only as an avocation. Business situations usually move slowly enough that there is time for briefings and conferences.” This statement is patently false today. You need only consider the abundant benefits that Internet companies like Google discovered by running online experiments and rapidly iterating their products to gain marketshare. (See A/B testing for a quick description of what this is and who is using it. Just read the course description; you don’t need to take the course!)
We live in a time of accelerating, technology-driven change. Good speech recognition technology took longer than Licklider’s expected 5 years, but it eventually came. And now, further significant improvements will rapidly appear. That’s a consequence of technological trends that lead to a doubling of transistor density, networking performance, and storage capacity across technology generations.
As we move to talking about the Internet in our present time, I hope we will pause along the way to notice the cultural, business, legal, and other shifts that took place in parallel with the growth of the Internet. And as we imagine the future, I hope we each try to overcome cultural norms and traditional ideas of our time.