Alan Kay at UCLA CS Department - Distinguished Lecturer Series (1993)

From Viewpoints Intelligent Archive
Revision as of 22:05, 5 December 2017 by Ohshima (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
hello my name is Leonard Kleinrock I'm
chair this time I step off into you we
have here a really exciting and dynamic environment is the constant spend time
with our faculty and student body each
year we select a few from among the very best researchers in the field and ask
them the high point of their visit in
their field they describe results the
open directions in which the field and
as you might expect these lectures always generate a great deal of enthusiasm and interaction
I'm really pleased you have chosen to join us today let's go inside the
lecture is about to begin
this is the fifth and last about distinguished lectures this year and I
can see you appreciate it is even enough cookies to go around so it must be a
success but of course was what got all of us here is our speaker today Alan Kay
who is a phenomenon and a legend is on time and I'm going to go through the
usual tell you somebody details personal things that motivated him in
his career and the things that drove him to do the things he did though most of
what you should be familiar with if you're not you're gonna find out the next 25 minutes is to make exclusive day
first of all let me say that there's two dominant characteristics that apply to
Alan and I think to many other successful people but Derek's very
clearly expressed in his career first is
the environment in which he grew up his
parental influence the things around him and the two great teachers he
encountered along the way and one great mentor as well but secondly it was the
environment of his professional colleagues and the the lucky events that
put him in the right place at the right time with an enormous amount of
potential creativity freedom colleagues in all of us it is your environment and
it is your background as well as your talent by the way that matter but really
all these other things that that influence how you use your talent
clearly his roots are indelibly printed in his mind his grandfather was an
author and he wrote over a hundred books and he was an organist today Alan is
two-thirds of the way through creating this enormous organ in his his intense
action at the moment grandfather confronts his father was a physiologist
the university professor he loved mathematics his mother was an artist and
a musician now here we see the Arts and Sciences in this background that have continued to
influence Allen and everything he's done in his life he believes that you can't
separate one from the other but the connection between the two is what make good systems come about Kenny Miller he
was born in Massachusetts in grade
school what grace he encountered his first
great breakthrough in the form of his teacher who taught fourth grade and had
a couple of tables on the side of the room with a pile of junk and some books
and she never said anything about what was going on when the kids were sort of
wander around occasionally poke around one day Allen poked around he found a book on electricity which explained how
you take a dry cell battery piece of
wire and then then when you create an electromagnet remember those things and
so a few hours later he was in class hiding behind this large english book
with little electricity book behind it and the parts in front of our lab made
this electromagnet and it picked up some paper clips and he let out a shriek and
totally disturb the class reaction a teacher was wonderful I just launched
into this great discussion about electromagnetism in electricity and got
the class involved in this thing which Alan had created I mean this was really Machiavelli she just laid these seeds
there for the kids to discover and it worked and that environment again
allowed him to open up and think about lack of rigidity a lot of flexibility in
the learning process PFC came into elementary school at the age of three he
learned to read and read hundreds of books before I get into elementary school this is no accident that he was
precocious at that point at any rate
this one was a true educator so much so he remembers that period his life
clearly as if a Technicolor movies there's no vagueness about going on in
that theory it's indelibly imprinted in the way he thinks
having read too many books before I got the grades goal he was also made aware that there was no
the concept of a single point of truth is false and that's a very difficult
burden for a young man to carry through our educational system his father said
listen ignore the garbage and get the a and know what you know don't listen to
what I say and it worked the flexibility
in that grade school environment the ability to create an environment and put
things together create what you want in fact found its way into the Mac user
interface think about the kinds of tools
you're given now it's up to you to create the sequences you want to carry
out your problem at Saturn at this grade school teacher ruined Allen until they
got to graduate schools all the other teachers the encounter between commit
yet another one so let's do it slowly he went to a junior high school in the 8th
grade he built the Tesla coil how many of you
know what a Tesla coil is how many you built a Tesla coil ok he's a wonderful
devices that create he's very high frequency very high voltage very long
arcs of electricity you can draw between 2 1
set it behind the wall in his teacher's classroom hey cuz he know this teacher
had a habit of leaning back in his chair
to the base of the skull totally that he
gonna flunk out right now we get the hell out of here anyway so he was right
on the verge of high schools we had a fun time instead of going to Bronx
Science like the rest of us guys did he went to Brooklyn Tech which is every bit as good and in fact it was there that he
learned that he didn't want to be an engineer okay
in fact it was being instructed by professors from NYU and Columbia
meanwhile he was into rock and roll in jazz bands he played the jazz guitar
professionally from the age of 17 to 27 when his family moved to high school
doing jazz guitar and girls and his GPA
followed well musicals for the theater
he then went on to University of Colorado in Boulder and set out to get a
bachelor's in two majors pure math and molecular biology which he did but
between sophomore junior years there was
an event that occurred in his life where there was and abuse based on for one of his
friends on an ethnic matter Alan took a strong position reacted and rebelled
strongly got dumped again and so he went
to teach banned in Denver was about to
retracted at the army opted to take the the test to go into the Air Force and he
did he got interested in computers decided to take the impossible IBM exam
which nobody ever pants become a programmer he passed it he went to
Randolph Air Force Base and programmed the great Burroughs be 5000 machine a
magnificent machine a very flexible architecture which influenced Allen
ludos internal structure and put these ideas to work in this in the small talk
environmentally they got involved with back in University Colorado who worked
at the National Center for Atmospheric Research on what he called Big Island he
simulated a CDC 6600 on a CDC 3600
Seymour the people who are supposed to bite the operating system that city 600
didn't do it so Seymour Cray did it and he wrote it he bought the entire
operating system in opto and punched the cards himself one error not bad
meanwhile Ala Moana disassembled to get source code out of that thing and
together they got the Chippewa operated system working and Ronnie and the Machine simulated itself instead of had
another machine simulator man only won three times its low simulated itself as
it could run to welcome nah he plan to go to graduate school and study
philosophy so he decided to shoots graduate school they also want to start
at the people science so how do you choose a computer science school he says well I want to choose the school that's
at least at the altitude of 4,000 feet so went to the Almanac and looked the
boiler schools that sounded spot that came and it was one University of Utah
he sent in one application to graduate school and went he went because the
second grade teacher he encountered someone named Dave Evans no should know
of Evans & Sutherland for example look
at his application threw away the GPA and the scores look at the fact that
Allen was into theater jazz guitar and other things and computer science and
hired him on a bed along with the other misfits that Dave Evans continued to
pull in every misfit was a great success he had a wonderful knack and a great
psychological bent Evans was the perfect match for Alvin
Dave started he would take a graduate and say I started with plus infinity and
it's up to you if you work the way down from there yeah ultimate confidence and
the people who work with it and they demonstrated a support they deserve that
confidence one day he walked into a
classroom and this big six-foot four professor followed them in and said my job today
is to disabuse you of any closely hold notions you came into this class with
and he spent to the best of that class doing that half the class dropped out
but those who made it through had a real mind-boggling experience and it
shattered some of the cross
he entered there in January 67 in graduate in summer 69 Dave Evans said
that two years of professional quality working at the PhD if you work on an
easy problem you gotta solve it fucking hard problem just welcome Alan worked on
hard problem namely the development of a personal computer and put together the
elements of machine following that had joined the faculty of Utah as well as a
Stanford University McCarthy's research laboratory following that he built pcs
instead of doing AI for McCarthy he was interested in Kitty computers and see
more packets were created tools helped children think along around then third
key figure in his life Bob Taylor who had been the promoter of the DARPA
environment that's the other environment Alan found himself in went to Utah
suddenly it was environment well Rob was saying do what you want to money at
three searches the researcher threw money at students and things happened
Bob Taylor was one of the key innovators about the concept I'm Taylor asked him
to join a graduate student a meeting only graduate students at summer 68 that
again continued Allen thinking about a notebook computer and created this this
content of a thinker toy in July 7 t0 after Palo Alto Research Center was
formed in September 70 Taylor moved there and immediately hired Alan as a
consultant meanwhile some small means like Alan Newell and Gordon Bell at
Carnegie Mellon yes Alan to join them he said yes Taylor said come work for me
at Parc Alan said yes teller and no to
Carnegie Mellon join Park spend ten years there in a wonderfully cooperative
environment again he developed the first raster character generator in January 72
did the bid mapping program which led to Windows in May 72 created musical
synthesis capability animation convinced Chuck Thacker at Parc to build the the
alto according to Alan spec Thacker took it on and
November 72 all the way to papal 73 it took him to build that machine three to
four months turn around the first of these machines meanwhile
ila have been putting small talk together on a bet then he poured it to
the alto Starkweather created the first laser printer there in 72 Bob Metcalfe
created even that there at the same time sanad a workstation object-oriented
programming laser capability Nathan F all with twenty people and the
environment that we still beneficiaries of today didn't evolve people from three
different labs who aren't cooperating had a sort of group where I get together and do it an Allen was the Pied Piper
seducing them all to make this thing happen he became a Xerox fellow he then went on
to Atari spent a couple of years there as the company went downhill at which
point a gentleman but they were Steve Jobs said coming work for Apple
we need your brains he's an Apple fellow he's interest in education computers
well what can you say after that
actually I got some interesting requests
about this talk one of the requests was
that I give a general talk instead of a technical talk one of the requests was
that it should have something to do about the future and the other one was
kind of a more vague one which is sort of well how do you go about doing this
stuff and I think the one thing we can take from Lenny's introduction is that
there's a there's a kind of a pleasant half hazardous when things are going
well that is the one way of thinking of
it is that the the turtle with a compass always beats the random rabbit you don't
want to have a strong plan when you
start I'll talk about this a little bit because there's there's a thing in
science and engineering schools where people are taught to be very rational
and to plan things out and have goals but in fact the stronger your goal is
the less chance you'll get a good one because if you think of it you haven't
done a lot of learning when you start a project the learning is yet to come if
you have a rigid goal a lot of the interesting things are going to get lost so we like to say that in research what
you want to have our directions and you think of the future as being some sort
of magnetic field emanating out there's a kind of a vision and what you are is
an iron filing lining up in that magnetic field but you don't know where it's coming from but it's sort of like
salmon going upstream they don't know where they're going but they can smell the gradient of the perfume that they
memorized when they were born up there and
just a few parts of few atoms in a in a million is enough for them to swim their
way up the gradient so let me just start
I actually of course have more the
insecurities of giving any kind of talks you always bring too much material
actually it's worse than that it
actually goes back further because there's a context that's been around for
a long time all the way back in 1962
here's what we had we had McCarthy who had already been talking about advise
taking agents for four years he wrote the advice taker paper in 1958 and was
one of the original stimuli and doing time sharing integrated circuits were
already about four years old so we could add those to look at mold
like process operating systems have been done on the fram the Atlas we're done at
MIT C TSS and the B 5,000 had one two
higher-level language computers will be five thousand was set up to execute the
semantics of alcohol there's already a personal computer as I'll show you
called the link that was done in 62 there's a network that was used to
connect cts s graphical user interface
sketchpad this is the 30th in honor of
sketchpad actually brought you along the first movie ever taken a sketchpad which is actually the summer of 62 it's always
instructional to look at something from the past and wonder if we've improved
much since then object-oriented design
well sketch pad was object-oriented I'll show you you'll see that quite quite
easily when we use it in the B 5,000 well it didn't have the concept of
object directly it was set up to be an object oriented architecture it's one of
those things where a really great system designer a bunch of insights about
software into the hardware of a machine in such a way that it was actually good
for a lot more than what it was originally intended for he actually put
in a few nice generalities that turned out to be very important later
hypermedia Engelbart got his first
funding in 1962 common sense reasoning McCarthy so the context that existed 21
years 31 years ago was quite remarkable now the other thing that was interesting
about that is that almost nobody paid
heed to that context that was the other
thing that and was we see it going side by side that there are these incredible
insights happening and then the larger
mass of people basically wearing about code I didn't I should have had another
cop column on all of the uninteresting things that people were getting
interested in like like IBM stupid architectures COBOL so
they're they're always always these two currents and in 72 by the time 72 had
come around there was gesture recognition we had the idea of notebook
computers speech recognition had already been done pretty well the ARPANET was in
full swing in fact there was already digital packet radio that some of us had
been fooling around with even the late 60s under ARPA bitmap display is
actually the overlapping window idea was already around by then and the first
complete object-oriented programming languages around so many of the things
that people are contending with in the commercial arena have been around and
the big problem that people have with the future is the problem that instead
predicting it in an interesting way what they want to do is predicted by
extrapolation and this is a lot because of course the interesting parts of the
future are things that are a little bit more like the weather they're not
extrapolate or E they're governed by physical principles and so when you look
at I mean one way one of the ways of thinking about this is you think of
several ways of inventing things one way is brainstorming which is the weakest
way of getting anything done there's this thing that a committee is the only
form of life with ten bellies and no brain
you get the square root of the combined IQ and in an MBA school and actually in
engineering schools they teach you to be gold or in this is find a need and fill
it going yeah let's find out what people
want and we'll build it for them that process successfully predicted
against almost all of the interesting technologies of this this century one of
the most famous is the story that everybody learned when we joined Xerox
which was how Xerox got going and it was written down in a book called my years
with Xerox the billions nobody wanted I told about how in the 50s the company
had gotten very interested tiny little company in Rochester got interested in
dry copying bought up patents did a lot
of work in fact they did so much work that when they finally came up with the
914 hardware they had run out of money
for marketing and building factories so it's a little embarrassing they had this
great invention can you imagine working for a company that spends that much for research
the reason was is the president of the company loved this more than anything
else and they just put all of their resources into this so I said what will
we do with this so they said let's take it down the road to IBM so they took it
down to I'd be on the ninth I think was 1958 they said the IBM here take this
machine we loved it so much we just want to see it go out in the world we'll just
charge you a small royalty just build it
and sell it IBM did what every large company does when they don't know what to do and that
is they went out and hired some consultants
the consultant is a person who knows a hundred ways to make love but doesn't have a girlfriend
so of course in America in America all
consulting companies are we have the first name of Arthur I was using that at
that time it's Arthur D little in Cambridge Mass and Arthur D little went
out and did the thing that they have been taught to do in business school
which is to go out and find out whether anybody wanted plain paper copying they
spent a year and a half compiling the six inch thick report that said no nobody's interested in plain
paper copying and there are two main reasons one is that nobody was copying
[Laughter] so one of these things if you only have
a straight ruler you extrapolate this by
the way was the reaction of the hewlett-packard marketing people to the
idea of the calculators nobody was calculating why build it the other
reason was the plain paper copying costs 10 times as much as mimeograph e I said
nobody will spend a nickel a copy to copy on playing bond paper so IBM turned
it down this long delay and the turn down so enraged the Xerox people that
they went to their board of directors and said let us use our life insurance
for security for loans for a factory to
build a new team the board said ok so the Morgans there themselves to the hilt
this is a 1959 and built the first factory for 914 and most of you weren't
alive but just a few years later they were the fastest growing company in the
world and less than 10 years in 1967 the
Federal Trade Commission said you are too successful we have to give up some
of your patent position thanks so much for all that effort then of course the
interesting capper to that story is just a few years later in 1976 Xerox turned
down personal computing
and personal computing last year was a 95 billion dollar a year business they
had a 10 year lead so click on these
things when we think about the computer is something more interesting strap
lighting from the present another way of thinking about it is one of the great
things about human languages you can lie with it all right think about it they
invent anything new because most of the
things we're doing today where the lies of a hundred years ago in 1895 Western
Union turned down investment in the telephone saying no sane person would
ever do business through such a contrivance the reason was they had a
printing Telegraph and left a trail a paper trail just like email does today
and they couldn't imagine that anybody would go back all this stuff going on in
London in England they turned it down for a completely different reason and
that is that because of partly because of the class system they actually added
eight posts a day in London and there
are books of correspondence so eight times a day there is a mail delivery and
so their correspondence of people who would have several exchanges of mail in
a given day you can tell from the thing and it's actually mentioned in the
Sherlock Holmes stories and so they could not imagine why anybody would want
to telephone because they're all of these messengers and put them out of work and so on so
so my theory on this stuff is you have
to some kind there are a couple of factors that help predict the future one
of them is in enormous amounts of romance something that this romantic
enough has an excellent chance of getting built because it's going to attract many people to it here's an
example of a romance that came about in 1945 in these illustration in the white
magazine version of the as we may think article by Vannevar Bush so one of the
things he predicted in the year was was dry photography and so he said sometime
in the future every house will have a desk in it and in this desk will be
about the contents of a small town library about 5,000 volumes of books
held optically and that there will be places for scanning these books and
there will be controls and we can point and we can call up information and we
can put in our own information even predicted that there would be a
profession that arose from this device which he called pathfinding so there'd
be people call pathfinders and what they would do is find interesting cross
trails through this information and they would sell the cross trails as
applications okay and it happened that
this idea was encountered by a number of
people in Iran did something about the part was in the Pacific he was in the
Navy in the Pacific and he read that when it actually came out I was only
five years old in 1945 but I discovered it in a science fiction book that I read
as a teenager buying a lot there that
when he handedly with lower case it was
a lowercase M emex he used it as a generic term and I knew that this
science fiction author had gotten that idea from somewhere else so I looked it
up and I discovered and I read that paper in around 1954 other people look like
this thing and Bush was President Roosevelt Science atomic bomb project
and he was also a professor at MIT so he was a guy or anybody could do anything
about it it gave a completely different way of
mine was something that Licklider used
as an example for what we should do with computers and people started doing stuff
about it here's a couple of early days
ivan sutherland - at the tx2 computer
much larger than this room Lenny and I both have fond memories of this machine
one of the computers run with 10% of its
wires disconnected she needs to use
transistor as one of your earliest machines to use machine by the way
that'd be equivalent to about CST about
maybe 480 k her cell but in its own time
it was incredible at the very same time that this is going on West was doing a
for biomedical researchers if we needed to do real-time experiment the idea was
not only will these people program but they're actually going to build it so he
actually had a summary and machines
can't even learn how to build their own and one of the design criteria on this
machine was that it be low enough so
that you could look over it they didn't want a machine looming over you when
you're using it so this is I hope you'll
be able to see this you could kill the lights
so here's everybody see it
individual that was the beauty of this
machine because Ivan could immediately get to work you'll see the one of the
important things a lot things particular
thesis and when you look at this display
just have to remember look at this thing
can it do and the answer was what it
could do it couldn't make good-looking pictures but it can simulate them alright so right away you had something
wasn't just a picture drawing facility you could actually put in a bridge and have a calculate the stresses and
strains you can put in an electric circuit and without explicitly preparing
it for anything about electricity just by specifying the constraint with
manufacturers now and moving around the
circuit to put in linkages and by under constraining the linkages it would animate through the languages
so we what he just he drew me in the
outline of this thing and he pointed to all of the different headings of the
thing he said to the system I want those all to be mutually perpendicular and
sketchpad just figured out how to do that so that dynamics to a set of
dynamics all right so this thing is not
just a graphic system but it's actually a problem-solving system about a third
of a mile on the side draws me and now
he's pointing to these two things he says make these guys parallel and now
the constraint is make these collinear
they're still there
they're very computers today that you can continuously zoom on a piece of
graphics guy goes to eat pot again wants
to make a rivet for this flange is he in
the idea of the rivet the society is
going to use that as a center for the arc he's going to draw here now again if
he points to the edges and says I want these to become usually perpendicular
further because it can do nonlinear
these guys to these guys
here's something it goes back to his
other picture he says okay give me a rivet
again he has a knob that he can rotate the thing
you can make it bigger and smaller there's gonna latch this rivet in there
my theory is looking at this program guy
says oh this crotch piece that's no good let me go back to the master drawing and
guess what these guys feel the change
all right so what we're seeing here are in strain the world's first graphic
system great it's one of the first dynamic
problem solving systems all in one year
and he said well I didn't know it was hard so when you look at sketchpad here
we see some
pressing ideas another ad by the way
we're i conce so the concerns in the later version of the thing were in the
form of icons iconic little structures that you could drop into a form so here
is these bounding a thing that's parallel and equal here an equal
constraint is gonna force the this figure to turn out object oriented it
has polymorphism for instance this this
word called display is what we call a selector and small talk and what's in
there in this instance structure is actually the subroutine entry for how
this particular kind of thing is is display see jewel attachment has an
inheritance hierarchy so here's his he has a type called variables see if I can
read any of these things various kinds of Polar's various kinds of constraints
various kinds of topological things and so forth it's also one of the very few
species on graphics in which the system was used to make all of its own drawings
so this is an incredible system and it's
one of those systems like Maxwell
this system was so fundamental to the
way we all thought that you can tell 40
year by a system that does what sketchpad
does yet today is impressive on its own
grant the only thing unimpressive about the video this now just kicks the shit
the average piece of software that you can go out and buy or even find at a
university reading it working on the
6605 because working on this huge tree
on machine
and what is to imagine a 6,600 sitting
on a desk no way he's 65 there's no way
I could imagine though I didn't even know of course I could see it was
exponential yeah I did 10 in 1965 I just
didn't have any overlap what I was doing
Gordon Moore actually kept on as it
wasn't about bipolar it was about a very
very slow way of doing silicon which is
an N Mo's field effect transistors of
course more these are predicting the view then it so one of the things that
Warren decided to do was to be a founder of Intel and they went out and did it
by the way this is what actually happened he said five doublings per
decade it's actually been heard a made
through 1995 understand you had a
blueprint of exactly what was going to happen to a computer years thing was
how long did IBM have to unbelievable
now by 1965 hey come up this now this is
a picture that was taken around with the
mouse in his hand and using a two-page display black on white this is what it
looked like this is what it looked like
from the top and he did group we're just
about close interaction it was about groups of people working we're gonna
that work with people down to Menlo Park
alright actually the father of the probe their various ways of looking to say Wes
Clark but you look at what we call personal computing today it is really
hard to escape the idea that angle Bart had this image now of course it was done
wrong was done on a time sharing system but very quickly we started thinking
about well of course time cycles we
better do something like this computer I did it appear this is a self-portrait
this is what looked like on its own display was a desktop computer for very
special desks this design was the
the thread was wire you wanted to fire
through the core and there's a nice exhibit in the computer Museum doing a
computer where you sequence required
ecology of some kind this tech this
machine used use a similar decided to
grind up some oatmeal and butter so you
decide to use some olive oil can't find any eggs and what you use for it so it
had this speckled was sort of had the right terms that was much better one
let's do it don't teach kids to drive when they're
five read and use media and clay and all
that stuff
pewter 'he's got me thinking about what if a computer were like a a book and so
in 1968 I made this cardboard because one of the
things that's interesting you know is my my desktop how much
but because of Moore's law think about
it as we could count make millions of these and therefore there is only one
Beast all doing a real personal
computing billions of things there are
many other metaphor stand themselves
agents agents
how to do it a island nation when you're
trying to figure out what you should work on important thing when you're in
badges it's a very similar one of my
biggest complaints about today is I think there are too many brilliant is
because the funders will only find trivial thing about the ARPA funders
absolutely didn't care what your some
money and take percentages
the way by brute force you basically
don't want to do research in the real world to come so you want to make
artificial and we made a whole bunch of them to try to differentiate them so for
example what drives this integrated
circuits what drives this is going to be pervasive networking so even if you live
long enough here's now the exact model
buzzes four pounds right this is
one of these things that go along with
it and so by making this artificial do here
told us what not to work on here see the
laser out here but this thing is gonna be no printing the structures that
you're gonna look for and they're also
gonna be dynamic all right so this has
to be
here's a really important which is all
this stuff easy to figure out this this
way is one of the fact around 1965
here's that machine's back thirty years
ago streams organized around
if your student reading temple versus
gothic cathedrals holding you can't
something very big it just collapses under its own way so the Gothic design
what if anybody here you know arch in
flying buttress
facing them right and at some point there was I thought what had happened
was that they wanted to hold the roof up the brace is steel
much stronger steel tation starts milk
there's a limit the matter
Oh is a clever way under tension and the
same amount of material
[Music] yeah and then there's this nice that we
since 1965 computer science has been about like these things
in gaming language they use why should
we be interested in this here's a very
naughty girls are very fragile and we
can grow a bit fertilize 50 cell
divisions of a factor of 10
of the Apple cells on the way because
there isn't actually enough genetic information for a baby I'll just give
you one little story how this stuff works
embryo logically the muscles are laid down the baby mama for they're wired up
by the nerves so how do the nerves find the muscles so you start going out from
the spine little amoeba things that are
there smelling their way along but there's not enough genetic to find the
right exactly the right area they go out the right will stay right and then with
the baby's exit the ones that are
working like 80%
- it'll never develop stereo because
division dynamic why you should be
thinking about biology here's a little e coli about one somebody has a better way
to do data structures
so over bunch of people who didn't have
to worry about AI data structures did
have to worry about way we're gonna be
able to do all
actually working on right really
speaking what thinking it is quite
remarkable how easy the next few years
way of putting things by the way but
think about it is thing that is being
broadcast broadcast everything that's
now being Russians are worth that's
by satellite the size of a few years
audience just like you these reels
because they're really
this is the film the film tracking shots
modern language
twenty years from now people will be
worrying about eager bodies cuff flex why because everybody graduate school is
worrying about
are talking now part isn't in it but it
is pewter is almost everything how to do
it first time by some students going
down and ask the question why is easier
to do diversity
running on unit from UNIX programming in
somebody else's programming language or operating system why not roll your it's
not that hard to do there do it but I
think friends has learn about this or
some similar language fairly early long and then another year later do a
processor for do both so the big
shortfall here is the ability of receivers to understand the implications
in the what America this
two peas in the middle of this thing said I rather be precisely these little
we'll get into if we think of Bill Gates
is the one-eyed person then he's definitely King and the person who's in
for the 2id person can actually see what
should be going on eating exercise
habits literally would rather die and do
yeah and we could say here we could say we
organized a diversity of our patients
for incrementals Jesus and so the
problem is that and then the actual
dispute is to cost and this Franois
filter syndrome so it's in you know when are you gonna pay the freight
- tough most Americans will say later and like IBM for instance is paying the
there's kind of McDonnell ISM just
people that what they sell is food are
really into this quickly from NSF right
they would much matter sure
then take a chance on doing something good the good things are actually harder
to do sometimes you lose on them we were
very fond add which is first discovered
basically the result is this and if you take the frogs natural food and paralyze
it with a little chloroform so the Flies are still alive you put them in front of
the frog starve to death they cannot see
the cardboard and eat it and we'll do it
over and over and over again eventually it will die and their various ways of
interpreting this but I like to think of these flies as ideas they're around us
all the time this is the thing that I think both Lenny and I would be very
strongly on and that is the most remarkable thing about the revolution
that's been going on is it was there for 30 years it was actually making choices about the
stuff was there those choices were pretty much made on aesthetic grounds rather than technical grounds deciding
oh yeah here's a fly what are where that came from home problem-solving interesting things
are things are connected to system don't
solve them you seem to watch to dealing
with something like an ecology or human body you don't fix human bodies adjust
all these systems that are trying to achieve homeostasis
all and reduce differences but back in science you look for the superhighway
we've learned 300 years in spite of any
reason to the contrary for highways
exist it is unbelievable I sign pointed
out there is most incomprehensible thing
because there's no a priori reason why we should be able to understand it to
the level that we do just leave you with
just one last set of thoughts here the
whole thing that's going on and in fact reason for our existence is because
we're in a place where our field
investigated we have a lot be very
different now so we're just a little
what happens over again is when things
get more interesting in it was found
that amazingly enough you could teach most of the population to read we're
gonna see a thing called health receipt in the next ten years because it's
obvious you don't want to wait until you get sick and then go to a professional
to help what you want to be able to do is to take charge of your own stuff and
a good example of that if I have a friend who's a diabetic and a couple of
years ago used to have to go to it now turns out for diet anybody hears it
turns out version of it possible until
almost too late to know where the blood sugar is what these poor people have to
do is go 24 hours find out if they were
in danger just a few years ago now they can go into a bathroom and thirty
seconds later they can find out exactly what their blood sugar this guy actually checks his blood sugar five times a day
he takes sugar and insulin and it's been a completely different life for this guy
he's he's taken the responsibility with
the help of professionals to do a lot of this stuff for himself and one of the
ways to think about democracy it's when the hackers don't become nearly as
important anymore but you extend the franchise over a much larger group and
if you think about a value system to what computer this should be about maybe that's one of
the ones that we should we should think about and here's you have people buzzing
you can dig a pit you can pile up stuff
to make pyramids very natural thing to do you want to make a geodesic dome you
don't do that with a pick and shovel you need something else what happens as
professionals get power tools is the
amateurs get them to alright there's no
place for the person with the show
for us
that's computer scientists the power tool this is where the science stuff
comes in what is the equivalent of Maxwell's equations well they're really
taught you don't see this as a metaphor
some people do who are theoretical academics but it doesn't actually get
through to most people who do programs and here's one way of thinking about
knack the equivalent of Maxwell's equations and computing the natural
science we're given the universe and our
job is to discover its laws God or
whatever but a computer science we have something maybe even more interesting is
we give laws that the computer and creates a universe all right so
Maxwell's equations are actually a synthetic synthesis of things rather
than analysis of things in the first version of Maxwell's equations came out
in here right now we can do better than
that now but to me the world of computer
computer dumb is divided up into two kinds of people there are people who
spent the two or three hours necessary to actually understand this in a very
deep way right because it's not the
Turing machine that's the important part well you can make it a pretty good
Turing machine computer with two instructions I think if subtract and
jump on neck less than zero but it's not a very interesting Turing machine
because you have to write gobs of code to do anything what's great about this
one is it has a slope like this right it's simple and what this says is it
doesn't matter what your programming language is
so anybody who understand this would never write in Pascal again right
because Pascal is a language in which somebody had the presumption to tell you
what features you were going to propane min this is how you're going to do
conditionals this is how you're going to do loops and by the way you can't do count ilysm and you can't do this and
you can't do that this says you can you
can do it in a very small space and to me the hardest thing to explain to
anybody about modern programming language design is that this aspect is
how you should criticize now anybody who understood this would never ever write a
line of code in C++ because it doesn't have this it's been left out and so what
people do is to write as though what's his name Strauss to actually knew what
he was doing there's no way he could know what he's doing none of us do so what you want is
a meta system that when your problem type comes up you can actually apply to
it this is what people should be learning does the modern versions of of
this in graduate school so that whenever
you go out and have to do something you're not bound by the existing tools you can always whip up a system that
will allow you to create the tools to make the tools that you need to do the
job so what lynnie actually asked me is
by having any words of advice in one
sentence or so and the thing I thought
of was something that Fineman said
science have to trust the experts
right you don't have to take things on faith at all and so my word of advice to
you is do not believe a single thing that I've told you