Stanford Artificial
Intelligence Laboratory (SAIL)
SAIL got
started in the semi-circular D.C Power Building, right center near Felt Lake
Artificial intelligence (AI) is a branch of
engineering devoted to creating intelligent machines, so I and others prefer to
call it machine intelligence or
machine learning. It is an ever-changing field in that as soon as a
certain problem is solved it becomes ordinary engineering and is no longer part
of machine intelligence. Here are photos of many
of the early participants.
SAIL
History. This
web site was created early in the new millennium to review accomplishments of
the ancient Stanford Artificial Intelligence Laboratory (SAIL), whose
government funding was initiated in 1965 by Professors John McCarthy and Edward
Feigenbaum. Lester Earnest was then recruited to design, set up, name and
manage that graduate research facility. It typically ran with a population of a
bit over 100 and nearly all participants seemed to have an enjoyable and
productive time there, though I had to deal with McCarthy’s financial
corruption by boxing him in. SAIL ran there for13 years but in 1977 was moved to
the newly renovated Margaret Jacks Hall in the Outer Quad of the main campus.
Along the way, I
invented a lot of stuff that came into use around the world as long as five
decades ago. I also initiated development of what became the first hand-eye-ear
robot with the help of new Professor Raj Reddy and his students. It took verbal
instructions on how to manipulate children’s blocks on a table and used
computer vision and a robot arm to do it – see Hear! Here!, a 1969 15-minute color video. I
also initiated the first attempt at a self-driving vehicle, the Stanford Cart, but McCarthy decided he wanted to
play with it and took it over, then terminated the project. However, a new PhD
student named Hans Moravec revived the vehicle and got it to navigate slowly
through a cluttered room, similar to research being done at nearby SRI
International. Raj Reddy and Hans Moravec then each migrated to Carnegie
Mellon University (CMU) and set up a Robotics Lab there, which again took up
the self-driving vehicle problem.
SAIL became a hotbed of
innovation that directly or indirectly produced dozens of commercial spinoffs,
some very successful such as Microsoft, Apple, Google, Sun Microsystems (now
part of Oracle), Cisco Systems (actually more of a rip-off), Facebook (a
totally corrupt spinoff) and the very successful Amazon (a spinoff from spinoff
D.E. Shaw & Associates). Many other successful companies were founded
directly or indirectly by people from SAIL including Sun Microsystems (later
purchased by Oracle), and Rambus.
SAIL also spun off many successful
academic research groups. For example, in addition to the CMU Robotics Lab,
Rodney Brooks earned his PhD at SAIL in 1981 then went to MIT and in 1997
formed the Computer Science and Artificial Intelligence Laboratory (CSAIL), an
indication that he liked the name SAIL. He also founded IRobot, maker of the
Roomba vacuum cleaner, and another robotics company.
In the late 1970s a SAIL group led by Lynn
Quam had undertaken a Mars research project in collaboration with astronomer Carl Sagan, who came by every few weeks to view
photos of that planet taken by satellite, looking for visible changes. Sagan later
put together the very popular PBS television series called Cosmos: A Personal Voyage.
SAIL staff member Whit Diffie initiated
the development of Public Key Cryptography in collaboration with Prof. Martin
Hellman, for which they were given ACM Turing Awards. A practical version of
that scheme was developed by RSA Corp., which was initiated by Ron Rivest, another PhD SAILor.
Given that the evolving CSD got distributed all over the campus, in the mid-1970s,
a decision was made to reconstruct an existing building on the Main Quad to
accommodate the entire department. I then
helped design what became Margaret Jacks Hall, next to the main campus entrance.
However, in 1979 as we moved in, the new CSD Chair chose to exclude one very important
part of SAIL, namely the Center for Computer Research in Music and Acoustics
(CCRMA), he had tried to get rid of them earlier, apparently because they were
mainly members of the Music Department, but I had succeeded in blocking him.
However, our move to Jacks Hall left them abandoned without a computer in our
former building, which was disintegrating. Fortunately, they were able to
slowly recover and are a thriving group today, now occupying the elegant former
residence of Stanford’s first President.
After moving into the new building, McCarthy shut down SAIL and fired
me, apparently to get out of being managed. I got several interesting job-offers
but decided to complete my education by becoming founding President of a
startup called Imagen Corporation in collaboration with new PhD Luis Trabb-Pardo, who had done much of the work on making a
small laser printer work for desktop publishing. I had acquired the printing
hardware from a Stanford grad who was the son of Canon Corporation’s Board
Chairman, so I then purchased a license from Stanford to use the technology
developed at SAIL and started off by seeking venture capital funding, mostly at
the companies on Sand Hill Road near Stanford but did not succeed. It took me
awhile to figure out that they all had lunch with each other, so if you failed
with a few you might as well forget about it. I even shaved off my hippie beard
and changed from a Hawaiian shirt to a coat-and-tie, but it was too late.
Meanwhile they were giving millions of dollars to “Me Too” disk-makers, all of
which went belly-up.
In late 1984 I was fired from Imagen by
another crook but by chance McCarthy simultaneously asked me to please come
back to Stanford, so I did and subsequently was appointed as Associate Chair of
CSD. However, I concurrently found myself in a mental fog, which I eventually
self-diagnosed as sleep apnea, but my Stanford doctor disagreed. I foolishly
believed her and I did not get it fixed until 1998, 14 years after the onset,
having spent all that time in a mental fog. Along the way, I decided to retire
in 1988, enabled by some lucky early investments that made me a
multimillionaire but not a billionaire.
In 2004, Sebastian Thrun,
who had earned a PhD at the CMU Robotics Lab working on self-driving vehicles,
joined the Stanford Computer Science Department, then located in Gates Hall,
and revived SAIL, which is still going today. Thrun
went on to fame as the creator of Stanley, a Stanford robot car that was the
first to win the DARPA Race Across the Desert in Southern California. He then
took that technology to Google, which much later put it into a company called
Waymo, now part of the Alphabet Corporation that includes Google and YouTube.
In 2009, I helped set up a reunion of
early SAILors to honor McCarthy’s passing and we were joined by the new group
of SAILors who were developing both self-driving vehicles and self-flying drones,
among other things. I then used this web site as well as a newly created Sailaway email list to invite participation. After that
success, I used both media to put together more reunions, the most recent being
in May 2015. I also found that posting opinions to the Sailaway
list often drew insightful comments from former colleagues, who are now spread
around the world in academic institutions, rich corporations, and governmental
bodies. I enjoy those exchanges, knowing the peculiarities of most of the
commenters, and plan to continue this practice until I croak.
In
2016, I decided to experiment with allowing anyone to send anything to the Sailaway list, which turned out to be a disaster, so it has
since been restored to a moderated e-list. Perhaps when I become disabled or
dead, someone else will to take over, but I plan to live until May 2043, at age
112.
SAIL
now has a lot to brag about, given that:
· The four richest publicly-traded
corporations in the world at the end of 2018 were all SAIL spinoffs: Apple,
Alphabet, Microsoft, and Amazon in that order.
· 18 SAILors have
received ACM Turing Awards, widely viewed as the Nobel Prize for computer science and
evidently more than any other lab in the world.
Archives
A
SAIL document and program archive running from 1972 to 1992 is available online
at www.saildart.org, courtesy of Bruce Baumgart <bgbaumgart at mac.com>.
The Old Sailors Spreadsheet lists people who
worked in SAIL or used its computer from its inception in mid-1966 until it was
shut down in 1992. The various columns, explained below, tell what they were
doing and provides contact information, which is mostly out of date, so I
invite updates and other corrections and request help in identifying some of
the project codes that I can’t remember. Please send to les at cs.stanford.edu.
The column headings on that spreadsheet
(Row 1) provide links to information about those columns as follows. In some
cases, there is more information in the column description, in which cases it
says [click for more].
ID lists individual
identifiers of up to three characters, letters or digits. [click for more]
First and Last Names use either formal
names or nicknames or both. People who won ACM Turing Awards, viewed by many
as the Nobel Prizes of computer science, are shown in red.
Projects shows the projects
this person worked on, but I don’t remember all these codes, so I invite
additions and corrections. [click for more]
SAIL
Years column shows
the years in which they used the SAIL computer.
Email
or Phone is for
contact information and those who have apparently passed away are marked
“Gone.” Most of this information is out of date, so please send updates to les@cs.stanford.edu. Some people would likely want to show
their Facebook addresses but because of my prejudices that will not be allowed,
since I plan to destroy them.
Web
or Address is for
that contact information, which is also mostly out of date. Send corrections to
les@cs.stanford.edu.
Innovations
L.
Earnest (ed.), J. McCarthy, E. Feigenbaum & J. Lederberg, The first ten years of
artificial intelligence research at Stanford, Stanford University Report No. STAN-CS-74-409,
July 1973. Summarizes research in computer vision and robotics (hand-eye
systems and a robot vehicle), speech recognition, heuristic programming,
representation theory, mathematical theory of computation, and modeling of
organic chemical processes, all performed under a contract with the Advance
Research Projects Agency (ARPA). Ancillary projects included the development of
a multi-processor timesharing system with display terminals on all desks,
advanced programming languages (LISP and SAIL), the first interactive computer
aided design system (SUDS) as well as research in higher mental functions,
computer generated music and Mars picture processing.
How a nosy bureaucrat
accidentally created the first social networking and
blogging service. Many people seem to
think that computerized social networking is a recent phenomenon, but it
actually blossomed first in1975 aided by a program called Finger that was
written for a different purpose, namely snooping on computer users. As
sometimes happens when computer programs get into the hands of users, they
flipped it over and used it for a different purpose, in this case for social
networking and blogging, though those two terms did not come into general use
until about 25 years later.
Choosing
an eye for a computer, Stanford Artificial Intelligence Lab. Memo AIM-51,
April 1967. Develops performance models for alternative visual sensors
and shows, among other things, that image dissector cameras (one of which had
been purchased at great expense by Stanford on advice of an MIT professor) have
much lower performance than inexpensive Vidicon
cameras.
Stanford Cart was born as a research
platform for studying the problem of controlling a Moon rover from Earth. It
then was reconfigured as an autonomous road vehicle for research in visual
navigation, then went into show business for a few years. It now resides in a
home for retired robots while awaiting a comeback.
J. McCarthy,
L. Earnest, D. Raj Reddy and P. Vicens, A computer
with hands, eyes and ears, AFIPS Vol. 33, (Proc. 1968 Fall Joint
Computer Conference), Thompson, Washington D.C. 1968. Describes
Stanford Artificial Intelligence Lab research facilities and accomplishments in
speech recognition, computer vision and robotics.
Choosing an
eye for a computer,
Stanford Artificial Intelligence Lab. Memo AIM-51, April 1967.
Develops performance models for alternative visual sensors and shows, among
other things, that image dissector cameras (one of which had been purchased by
SAIL at great expense) have much lower performance than inexpensive Vidicon cameras.
A hummingbird with
range,
2009.01.01. The radar atop Mt. Umunhum, south of San Jose, California, which
was part of the SAGE air defense system, managed to get even with me in 1966
for badmouthing the crooked system in which it operated.
SAIL Away, The Analytical Engine, May 1995. Reviews some
spin-offs of the Stanford Artificial Intelligence Lab (SAIL) that helped
populate Silicon Valley.
J. McCarthy & L. Earnest, DIALNET and home computers, Proc. First
West Coast Computer Faire, San Francisco, April 1977. Described a
system that provided ARPANET-like services to multiple users via switched
telephone circuits, including email, file transfer and remote login.
J. McCarthy, D. Brian, G. Feldman, J. Allen, Thor—a display
based time sharing system, Proc. SJCC 1967. Describes one of the
earliest display based timesharing system.
Oscar Schwartz, Untold History
of AI, a series of articles from IEEE Spectrum.
Having Fun
C. Rieger, 54 second video of a SAIL Volleyball Game
from1972
featuring Russ
Tayler, John McCarthy, Norm Briggs, Les Earnest, Dave Smith and others.