‘John McCarthy’ awards showcase Stanford’s ‘nothing is impossible’ culture in artificial intelligence

The first-ever reunion of Stanford’s Artificial Intelligence Lab highlights innovations both prominent and unorthodox, including one very special vending machine.

By David Murphy

Stanford Artificial Intelligence Lab (SAIL) members reunited Nov. 21-22 to honor many accomplishments—Stanley, the autonomous robot car that won the 2005 DARPA Grand Challenge, the development of an influential robot arm, and an early music synthesizer—but they saved their biggest applause for mention of their beloved computer-controlled vending machine. By refitting and programming the “Prancing Pony,” as they called it, back in 1972 they transformed an analog vending machine into a digital accounting system that tied purchases directly to a hungry user’s tab.

But if sheer frequency of mention is any indication, then the historic gathering of more than one hundred SAIL alumni, faculty, students and staff spanning all timeframes of the program’s existence since it began in 1966 was also a tribute to co-founder and oft-billed “father of Artificial Intelligence” John McCarthy. For not only was the attending 82-year-old Professor Emeritus often cited as a source of inspiration by the twenty SAIL alumni honored at the gathering, but the new medals recognizing SAIL’s best projects bear his very name: the John McCarthy Award for Excellence in Research and Research Environments.
“This is a big reward for having fun, which is all we did in those days,” said Phil Petit, who received a John McCarthy award for his work in developing the first interactive electronic design system in the late 1960s. The Stanford University Drawing System, also known as SUDS, allowed designers to fully automate the logic and physical construction for electronic circuits.

In their attempts to find creative solutions to problems, even ones that seemed outright whimsical, SAIL alumni built a window to the future of computing without even realizing how widespread their technological achievements would become more than forty years after the fact.

Petit was also a driving force behind another SAIL project that wasn’t so much known for its nobility as its spirit. He takes credit for spearheading the carpentry behind the official SAIL attic, an eight-by-twelve-foot sleeping area that helped solidify the “home away from home” feeling of SAIL headquarters—the D.C. Power building that originally housed the lab, some five miles southeast of campus proper. The lab is now on the first and second floors of the Gates computer science building.

“We used to work mostly at night in the wee hours because that’s when you could get [computing] cycles. It seemed more efficient to sleep at the lab than to spend all that time going home and coming back, so I decided we needed an attic,” Petit said. “We took the ceiling tiles down in one of our offices, ordered some lumber, and got up there and built a platform. There were a couple of cots and sleeping bags.”

This sense of spontaneity—coupled with the SAIL belief that “nothing is impossible” as described by fellow John McCarthy award recipient, and former Professor Bruce Buchanan—led to the growth of numerous creative customizations and tweaks designed to improve daily SAIL life. Some of these were strictly for SAIL’s own benefit (and perhaps amusement), like the Prancing Pony, or a radio-controlled channel selection device for SAIL’s television.

 “For all the diversions, it was a supportive, friendly environment in which the super-hackers, or wizards, willingly helped the rest of us understand how to make the system work,” said Buchanan, now a professor at the University of Pittsburgh.

Building the future of computing
Today SAIL-affiliated professors number 15 of the 34 current computer science faculty positions at Stanford. Of the 55 recipients of computing’s highest honor, the Turing Award of the Association for Computing Machinery, 16 are connected to SAIL in some capacity.

John Chowning received a McCarthy for developing a method of digitally synthesizing the sounds of all known musical instruments in 1967, which eventually became one of Stanford’s most lucrative patents. Vic Scheinman won for developing the Stanford Arm in 1969, the ancestor of modern industrial robot arms. Other honored innovations included influential programming languages, and a pioneering videogame called “Spacewar.” Several current and recent lab members working in the research group of SAIL director Sebastian Thrun won McCarthy awards for developing Stanley which drove more than 130 miles through the Nevada desert without any human intervention to win its 2005 contest.

In 1973, Martin Frost programmed a digital news platform allowed SAIL users to receive news updates from both the Associated Press and New York Times wire feeds at the top of their terminal screens, which they could search through or customize based on keywords of interests. This grandfather to Google News became one of SAIL’s most popular programs and earned Frost, now a systems manager for the computer science department, a McCarthy award.

Senior Research Scientist Emeritus Les Earnest, SAIL’s Executive Officer from its inception through 1980, was tasked with keeping tabs on the various people and programs running throughout the 24-hour SAIL cycle. He also served on a committee of researchers who helped launch the ARPAnet, which became the Internet. As a result, SAIL has been on the network since 1971.

“I occasionally lived 25 hour-days,” he said. “Each day I would rotate ahead one hour so that over time, I would intersect with everyone, although it got to be kind of strange when I would have dinner at what was most people’s breakfast time.”

His John McCarthy award-winning solution to this issue, developed in the early 1970s, was a program called FINGER that allowed users to check and see who was currently logged into a given system, how long their system had been idle (to help determine whether users might be at their desks) and, if not online, the time of their last access. Earnest later built a means for attaching text-file updates which allowed users to update others on the status of their projects or future vacation plans.

“This whole thing then sort of morphed,” Earnest said. “It thus accidentally became the first social networking program. Those people … could use the plan file to give personal statements of various sorts. In other words, it became a blogging service, and it was widely used for that long before the term blog was invented.”

David Murphy is a Contributing Editor at PC World and weekly columnist for Maximum PC

Supplemental video: SAIL programmer Georgia Sutherland talks about the early days

We are interested in your nostalgic photos and the stories they tell. If you'd like to share them with the Stanford Engineering community, e-mail them to David Orenstein , Manager, Communications and P.R.

– December 2009

Alumni comments:

Drew Lanza (BS, MS EE 1979)

I regretted not being able to make the reunion and dinner.
There is another side to Professor McCarthy and the SAIL boys and girls that should get written into the history books.

When I arrived as a freshman at Stanford in the Fall of 1974, all the undergraduate computing courses were taught using punched cards and the cards were read and run on an IBM 360 mainframe. I knew that other Universities like MIT and Princeton (famous for inventing the language BASIC) had timesharing systems they used to teach their students programming. Students could sit at terminals and program away.

I complained loudly to some of my Professors and I believe that it was Professor Tuttle of the Electrical Engineering Department who steered me towards the Committee on Computational Facilities, a recently formed group that was looking for a student representative. A fellow undergraduate, Anne Bucksbaum (co-incidentally now married to Tom Friedman of "The World is Flat" fame), goaded me into 'stopping all my whining and fixing the damn problem'.

I became the student representative to the group. I cannot remember if Professor McCarthy was an actual member of the Committee or not, but he stood out to me as a man who cared about undergraduates and cared about them receiving the best education and facilities he could help to muster. He was a huge inspiration to me and gave me the confidence to push the deans hard to fund an undergraduate timesharing facility. He always made time for me and was generous in sharing his understanding of the Byzantine politics at Stanford. I met with him numerous times at the SAIL lab seeking his advice. That was my introduction to SAIL.

The rest, as they say, is history. Once the deans had agreed, Professor McCarthy did another very generous and smart thing. He 'gave up' Ralph Gorin, a very talented sysadmin, to move down to campus and run the new LOTS (Low Overhead Timesharing System) DEC computer systems. It was a huge success and vaulted Stanford into a leadership position among peer Universities.

Along the way I took classes from and got to know most of the SAIL people. There is a tendency to think of them as trolls living up in the foothills at the lab off Arastradero Road. Nothing could be further from the truth. They were vital parts of the main campus, too. They invited this geeky undergraduate to many of their bashes and were generous with their time and spirit inside and outside of the classroom. They loved what they did and wanted to share it with and inspire everyone around them.

I ended up majoring in an offshoot of AI, because of their influence. I received a BS and MS from the EE Department in Adaptive Systems Programming and Control - the 'hardware' side of AI - from Professor Widrow in 1979. I went on to teach a class in the subject (EE274), start four companies here in the Bay Area, and am now a successful Venture Capitalist.

My willingness to explore the unknown and to seek out on my own were inspired at Stanford and fostered by great men like Professor McCarthy.