Alan Kay, 2015: Power of Simplicity

From Viewpoints Intelligent Archive
Revision as of 03:04, 21 November 2021 by Ohshima (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
I'm not a vendor
I'm not trying to sell
you something for money but I'm
going to put a few ideas in front of you and
I view my job here
is first to get us to
lunch in a reasonable time
second
to get questions from you some
of these will do at lunch but I'm going
to go against my the people
invited me here and I'm still going to go solicit questions
at any time
because otherwise I
could have just made a video and
we could do some sort of conference
call for the for the questions that's not what I
want and then the the third part is
to put a few ideas in front of you
that might get the questions to be
important for you I
was asked to talk about simplicity
and there was some interesting paradoxes
there but the first I like to
deal with a metaphor about ideas
and the question is our idea is
made out of ponderable matter in which
case ideas
oppose each other they can't be in the same place at
the same time or our idea is
made out of light in which case
we can have any number of ideas in the same place
at the same time we don't have to choose between them
and sometimes even the color changes can
be very suggestible similarly
our ideas things or are they processes so I'm going
to take the right-hand side view
here and put a bunch
of things in front of you some of which
may seem somewhat paradoxical in
fact the likes
curve here this is a curve everybody knows in
this room represents quite a bit
of the last maybe a hundred years but
certainly for most of you the time
that you've been in your
career there's always some exponential
happening and at
any given point in time represented by the
vertical line there there's a likes factor
in corporations it's
not just the competitive pressure
but all the legacy
this is in that's that
need to be tended to people think but
are costing more and more fact
they're costing so much in the realm of software
that corporations have actually been
unwittingly destroying the
very agencies that could help them namely corporations
are telling universities that they
need people who are experts
in programming in this language or that language all
of those languages are completely obsolete but
hey happen to be the ones that your legacy software is written in and
because universities have decided to turn themselves
into businesses they're feeling
the pressure from businesses
because they want to get money from the businesses and so
the businesses are actually undermining their own future
because you're systematically
killing off the people who might actually
come up with much better software solutions number
fifty sixty years ago when I started most
corporate software is written in machine code
somebody had to invent the higher-level languages
that were good for a few years
unfortunately we're still using them ones
basically from the 60s things
have to be reinvented over and over again and the
other thing about the Ickes curve is
the actual complexity part
of the Ickes might be a lot lower
might actually be
a tiny part of the Ickes
and perhaps we could call the the
other part of it from that nice
little thing down there up to where
the Ickes is we could call that complications so
complications are basically noise
bad
technique old technique
human bumbling
inability to get on learning
curves you name it's a whole bunch of factors
and quite a bit of the
stuff in corporations today if
you examine it I think you'll all agree with me is most
of you or CIOs if you've looked actually
at the code and I think the tide is turning with
CIOs because when I first started giving
talks like this to CIO is about eighty percent
hem were from the financial organization they really didn't
know anything about computing and so it's very hard
to explain what computers were possible what what
they could do but now it's it's different
you've looked at the code in your company
you'll realize that wow I've got millions of
millions of lines of code there and I have
more than a sneaking suspicion that a
lot of that code is actually in my way it
doesn't represent the actual bang
per line of code that we'd expect
from a higher-level language so going
back in history anybody
ever seen a diagram like this before
somebody they don't show
them in business school but the other
part of the campus so this is a
this
is what the planets the
paths of the planets travel
over an entire year this was actually
done in the 17th century and I'll show you shortly one
done a little bit earlier and you can see why they're
called planets the planet
is a Greek word meaning wanderer so
these curlicues are actually for instance what the orbit
what mercury does if you look at it every night it
will come into view and it will go this way across the
sky but then all of a sudden it will come back going
in the opposite direction you can see why
hard to do astronomy if you have two moving
things in orbits going at different speeds
and viewing each other at different times of the year
very very hard to understand what's actually going
on and there were some religious beliefs going
all the way back to the Greeks that said God is
a perfect being in the air four would
only use circles to
explain the course of these orbits
and it turns out the
orbits are not circular and
especially if you put the earth in the center instead of the
Sun and so they came up with this idea that
you see here on the
top right which is okay we'll
take a circular orbit and we'll put another circle
so as the planet
goes around in the larger circle it's
also going in the smaller circle and
combining the two circles as you see there will give
us something that has these loops that we see this is
called an epicyclic theory of
orbits and when
Copernicus went over to putting the Sun
in the center he also believed that God
was perfect and so he decided they would be circular orbit so
comparing whatever you've heard in school was wrong
Copernicus's scheme only made things
slightly more simple but in fact it still had
this mess and
you can put more emphasize
just sound like anything you're familiar with
you hang on to that old theory no
matter what it is and you start putting fixes in just
like software let's not rewrite it
oh no let's not find a better way of rewriting this
awful oh no let's not do that let's just patch it
natural human tendency that's what they
did in astronomy back then this
is before science got invented for real
Tycho was extremely
meticulous and
Kepler the site
who worked with him decided to believe his measurements
this is for Mars and
Kepler decided well
I sort of believe in God but let's try something else
so the first thing he tried were ovals
because they would fit the actual orbits
better and not quite enough then finally got right he
said what about ellipses and he had thought about ellipses
because as the next thing after circles in it
he didn't try ellipses for years you know why because
he figured that the people before him who were
really smart had already tried ellipses and found
them wanting no
they were really smart but they were too
dumb to get off their circles that they loved
so much so when Kepler plugged ellipses
in lo and behold everything
cleared up
even comets were explained
bingo what's
the expense of getting
simplicity and I'm going
to go over this a few more times because it's it's
not the only way of getting simplicity but boy one
of the things that's worked the best the last three or four hundred years
is you get simplicity by
finding a slightly more sophisticated building block
to build your theories out of its
when you go for a simple building block that anybody
understand through common sense that is when you start screwing
yourself right and left because
it just might not be able to ramify through the degrees of freedom
and scaling you have to go through and
it's this inability to fix
the building blocks that is one of the largest problems
that computing has today in large organizations
people just won't do it
okay
why won't we do it well one
things is that our brains were set up for
dealing with about a hundred people at a time living
by our wits hunting and gathering and
dying in the same
were born into for hundreds of thousands of years
there's no concept of progress
in our genes we
just don't have it but like all animals
we have an enormous set of genetic
apparatus to make us good copers
anything happens
to us we can find a way of being resilient about
and adapting to it we're copers and adapters
and so when we come up against difficulties
our tendency is to
cope with these difficulties
it's like working for a company go into a company and
the company seems
sort of screwed up maybe
you can quit you
can cope but your chances of actually
changing the company are very low because nobody will listen to reason
right that is not what
the company is there for they are there
for their a task this
engelbart the inventor of the mouse pointed out years
ago that companies are devoted to their area a task which
is what they think they were about most companies
do not have a very good be process which is supposed
to look
at the a tasks and make them more efficient but
almost no companies have a see process
which questions the tasks are our
goals still reasonable our processes still reasonable
that's the last thing it gets question
because Wow how do you do will change if
we're going to change our basic process in the midst of
hammering on us for quarterly earnings so
this is a huge huge problem and
yet it can be done it's just really one
really sees it so here's an old
model from the 19th century of
memory which
actually in the 21st century
has come back as a pretty good one as a metaphor
anyway so the idea is that rain comes
down on the ground and there's a little regularities
randomly there and at some point
those regularities will be a little more
responsive to the rain and a little channel
will form the channel acts as an amplifier and so
that channel got started it starts
funneling lots more water through it other
water is draining into it and all of a sudden it starts cutting deeper
and you
get these gullies and
you get down into these gullies you have to
remember to look up because everything down there in this gully is
kind of pink you
can think that the world is pink and in fact
if you get into a real gully one of my favorites is
Grand Canyon by the
way that's only a hundred million years of erosion to get
the Grand Canyon it's relatively recent get
into one of these things and the
enormity of what you see outwards
Dwarfs what you
can see if you look up if you've ever been on one of these
things you're just in a different world it's
a pink world you
don't think about climbing out of it you think
about moving along in it and
so I'm gonna take that gully
world and flatten it out here's our
pink world and let's take human thought as being
like an ant and that ant
can move all over this two-dimensional
world our world to us is basically two-dimensional
maybe it's fear to
larger beings but for us it's basically flat
and we can move all over it we can make
ncounter obstacle fools we
can solve those problems and get around them so
in this two-dimensional world here we have all the
paraphernalia of living and thinking
and if we grow up in that world we
don't know it's pink right
because that's all there is that
is the background color it's the thing
we are least interested in because it's the most constant thing
but every once in a while
might have a little
blue thought could
be waking up in the morning taking a shower but
then
we grew up in this world we went to church
we have parents going to school
pink
pink is what the reality is but
every once
in a while you get a cop out
at kerpow is out of
that world it's actually an escape
from that world and the
old days when people had one of these they would start a
new religion because
if you've how many people have had a Kirpal
of any kind like
the technical word for kerpow
is holy how many people have had it come
on how many people have had a holy holy
where does that where does it
from so the subjective sense we have
is we didn't have that thought
something put that thought
into our head it just happened so
of course you don't have science
and you're not wired to check out
the kerpow seems to come from the
heavens okay so
we dip into another world
this world let's call it a blue world blue
plane world and there are three things
here this explains
why we have trouble making progress if
we treat our beliefs as
reality than
how sane is blue
the answer is well it's not sane
sanity is relative to the things
we believe are true so
the first thing is you've immediately turned yourself
into a crackpot for a few nanoseconds that's
one of the fists that comes down and squashes you back out into
the pink well I don't want to be crazy
second one is
when you're trying to explain this idea to somebody else
they really have to go through a similar process this
is probably the most difficult thing
about an age of invention like the one that we live in the
inventors actually invention is actually
relatively easy with the right kind of funding
so the problem
getting
something is essentially
pulling other people into a blue world given
that the blue world isn't a complete
isn't really completely nutty and
the third idea is that the blue plane is also a gully
so they have a half life to
each one of these things
salvation 20
years later it's got
its the albatross around your neck right
and so anytime a company does something
successful and you can talk to
about it they really think it's like they invented
something really important no
in fact they just found a heuristic that's working for
a while and if they forget to re-examine that heuristic
they're going to be in the same plight once once again
so this is a real
picture it's not the sign
says do not touch any of these wires
and it's important to realize that
every single one of these wires was
a solution to a perceived problem
there's no other reason why was done it was done over
a period of time
this is related to this idea no
anybody can make a doghouse can make it
out of almost anything matchsticks even you can
make it out of cardboard you can make it out of just
about anything maybe toothpicks you
might have to take some care nothing
to it but let's try and just
scale that doghouse by a factor of 100
so now it's about 150
feet high it's tiny
compared to the Super Dome but that doghouse will just fall
in on itself completely has
no structural integrity and
the reason is that when
you double a solid the
mass goes up by a factor of eight and the
strength in simple
materials like wood and beams and stuff has
to do with the area it's like the strength and our muscles this
is why Jim Nast's are small they're small
because they can have relatively
large muscles they have short muscle arms and they
weigh quite a bit less that's why a grasshopper can
jump a hundred times its own length and we can't
they have the same kind of muscle fibers as
we do and so the scaling thing
takes what is a very nice idea for a
dog and one you can have in two seconds and hit it together
into something that you really do not want
to carry into any kind of larger scales
because there there's no connection to the
scaling and if we come back to what
we can do with out
special knowledge we wind up with
pyramid it's the only big thing you can build without
knowing how to build which is just a big garbage
dump and plastered over with limestone so it looks good
but if you think about it camp
it has no room inside so
superdome you have to do that other thing you have to
go back to a different conception of what the materials
are was you're actually tensile structures
and then you can build enormous
domed structures that scale
very very well so if we come back to
this tangle I just
put software in here it could be anything
the result of incremental problem-solving
so
this research community I came out of Vance
Research Projects Agency in the 60s and then Xerox
PARC which was an outgrowth of it
basically it was a bunch of
small number of people who have big ideas
who did not have big resources
they didn't want to give up their ideas and
so they were faced with this dilemma is that they could
not handle they wanted
network that would go over the entire world it was called the intergalactic
Network back before it was called the Internet and
they could not use
any technique that
Bell Telephone 80 used
because didn't scale
just completely out of the scope and
finally in the
60s one of the the organizing
insights
was hey computers are virtualizes that's
what Universal turning machine means what
that means is forget about wires
you don't need no stinking wires
what we need to do is
to understand how to organize systems as virtual
entities and we can render some
and we'll render some of them in software but
in fact everything winds up being something like a single
communications
line with an arbitrary number of entities
on everyone can talk to everyone else
and all of a sudden you've thrown away
all the things that Bell Telephone had and every
piece of the way most offer was done and replaced
it with a simple messaging system all
of a sudden few people could do amazing
things so
this is an example I'll
bring this up again a couple of times but
basically need to solve the context need to solve the Grand
Canyon problem most people
are rewarded in school for solving problems when was the last
time your child or you were rewarded for
finding a problem
you found a new problem we've got too many already
hi
where's in fact finding
problem is is the big deal and people
fight you every step of the way they'll
fight your kids in school every step of the way if they're a
problem find your type don't let the teachers hurt
them
most problems are bogus because they come out of
current context we're trying to get beyond the current context so forget
about problem solving it's just a bad heuristic
it's the last thing you do and so
you get these leaps so
here's a leap out of the context of the
20th century mid 20th century which
is a gear kind of thing everything is closely articulated
the interfaces are very tightly
bound etc etc the
problem is you can only make a thing
with about a thousand gears in it before it seizes up you
get the tolerance is good enough that sound
like software
just can't do it but when you go to biology
we have a hundred trillion cells
in our body each
every body in here has about 100 trillion cells ok
I'm going to ask the audience a question you
haven't asked me one yet but
the question I'll ask is who
knows how many of these hundred trillion cells in
our body have our DNA in them
okay
predominant answers all of them anybody
else got an opinion
so it turns out only 10% of them
only ten
of those hundred trillion cells have our DNA the
other ninety trillion nine tenths
of the cells in our body are slide
and that
slime has
people have been counting the species of
microorganisms you know bacteria e.coli
s one of them that should have been a clue Iko
coli has its own DNA which
is very different from ours
so late last count I went
on the internet the other night the sea was getting
close to 25 thousand different species of
microorganisms most of which we have no idea what
they're doing inside of us and they're
about 1 mm the size of our regular cells so the
90 trillion cells of slime is about
a basketball-sized of stuff all
over us everywhere that'll
make you feel better about lunch
but the
point is nobody's come close to building anything
with a hundred trillion parts or a trillion parts
without we
don't the only things we know that actually work with that many
parts are biological things something
to think about and that gave us
researchers back then a kind
of a unified vision everything
this is a self-portrait of the Internet after
we built it but this is the image
hey everything is like this this
is a biological model weak we can't scale
and have central control big
problem with companies
they start off like families with a head of the
family try and get bigger is why monarchies
are tough right no
way you have to find a
way of distributing control and distributing responsibility
in an ecological way this is
not thing things that human beings like to think about people
are uncomfortable who's running the show people
say well the answer is the Internet does
not have any Center and it's grown by almost ten
orders of magnitude now without ever breaking your
software breaks all the time the internet has
never broken it's replaced
all of its atoms and all its bits at least twice since
it started in 1969 it has never been
taken down for maintenance think about
that your software could be like that the
software we did at Xerox PARC was like that your
software could be just running eternally
and so everything
at the expense of going
to something more complicated than a data structure or
some wires everything can
be built out of a single kind of entity that
has functionality inside it provides services
on the outside and there is something
like a cell boundary on it it's worthwhile
thinking about that
so if we come to Park
here a couple of things we did personal
computer this is basically in the 70s bitmap
screens the GUI WYSIWYG
and desktop publishing what we like to call real
loop now since the term object-oriented got taken
from us by C++ and Java
laser printer PostScript
Ethernet pierre-pierre
and client-server and about half of the internet
we had our own internet so these are about nine and a
half inventions and how are
they done and who did them well
25 researchers did all of them
25
think about it
cost about 12 million
dollars a year in today's money every
single company in this room every single 500
fortune company these are fingernail clippings on
your IT budgets you waste
more than this every other week and
espite that there's not a single company in America
that until
recently that has even taken the venture of doing
a process like this you have to ask
yourself why it's a question
you really wouldn't need to understand because we're not talking
about money here return 30
plus trillion dollars in counting actually around 35 trillion at
last count what
was the problem and
was not the
problem of Xerox not making any money this is a
story made up by companies to avoid having
to contemplate doing a long term Research Center this is
an urban legend just absolutely untrue in fact
Xerox paid for all of Park more
than 200 times over with the laser printer
alone isn't that the
most obvious thing how many billions have been a big
bottom with Xerox is they only wanted to make billions
and that's the problem with most companies
because when you're doing this kind of
you're actually in the trillion dollar range and
no company has ever been able to step up to the plate
and just one
other point because I'll get to it in a second
again is that of all of these inventions
we had to have all of them you
had to do all of them but the one that
reached this stuff out to everybody
was probably the GUI because it
is the meeting ground between
people who don't know computers speak and what
the computer can do so
it is the thing that allowed this to go out to multiple
billions of people down
ok so now you guys
are mostly CIO so I don't ask you this question but
what I mean CEO I
always ask them are you
be in business in ten years and prospering what
do you think they say
they look kind of like that
you know I real I was
searching for a face then I realized wait a minute Cheney Cheney
has got the perfect face
for this guy
and my next question is well what
your ten year plan and the reaction I get is that
right
think about it the
idea of a ten year plan that people are serious
about is just it's
fake companies just don't
have it they don't set themselves up to
be able to deal with this thing which
is really just to find hope that they're going to be in business in
ten years they have no idea so
let me ask you a question just
think for a second where were you
ten years ago ten years ago
it was 2005 country
was going through some real problems
back then and
then realize how Fargo
does 2005 scene well it doesn't seem that far ago
but it was ten years ago so ten years ago today
was ten years in the future think of what we could have done
if we didn't think ten years was big
we could have thought all kinds
of things in 2005 and pulled them off by now
but because ten years seems impossibly
long in business terms most of these things never even
get talked about way too far off we're
worrying about next quarter
so if we plot out an invention
process like Xerox PARC
let's imagine we can come up with a 10-year vision
and
basically everything we did at Xerox PARC was thought
of as a five-year horizon when we did something
it turns out five-year horizons are necessary
in order to get done earlier
so in a five-year Rison most
come up out in the first three years if you set a three year
horizon you're not going to get them
because death just isn't the way people work that
five-year horizon allows people to do the right thing
the first year if
narrow it in too much they will not do the right thing than
the first year and same
thing innovation taking an idea
out into the marketplace five-year
horizon there's a transfer process
most of the when I was at Apple most of
the innovation processes for big things we did
took about three years but were organized
kind of like this and so
get when everything's going well the wind is right the
Creek hasn't risen you get about
a seven-year thing out of this ten year framework
that you have to set up well that's kind of interesting
so suppose we
had done this we go back seven
years now this is 2008 that was just next door
and what
happens well same thing can
your vision all that stuff I showed before seven
years bingo today is the day
that the seven-year thing
came out and if you just study how this stuff works from
things that are basically new not simple
increments like new web apps but
things that are new seven years is
about the fastest you can do it you can almost always do it in
under ten so that means a
small amount of money but
allocated over a time that could be longer
than most CEOs stay around is
actually this is the problem
in government it's hard to carry out long
range policies when the politics are changing all
the time now the problem with this
in America is there's no business reward system
here because the costs for this have to be expense
right Reuben it's like the
Reuben is
a good guy here nobody has been more
delightfully clever at helping these
good processes along than Reuben in my experience
but the cards are stacked against
companies because
every dollar you take out of this thing is
a dollar that could improve the bottom line for this quarter
reporting and that is wrong
change that law you've got every other law changed
look at the laws on depreciation
crying out loud those
are ridiculous but they're in business's favor you
have to have something that
contrasts favorably rather
than the huge disparity between this where you've
got the right process going but it's
hard to pay for it with an idea that is
not so good which is to
acquire every time you want something right
acquiring acquisitions you can do with different money but
you start killing your corporate culture
right you have that problem but
yet it's much more favorable because of the kind of
you're allowed to use for it so this is crazy and
until that happens
the universities are the only place that are
going to save you that is how this stuff got here ARPA
funded universities because
if I BM couldn't do it they
were spending billions literally billions of dollars
in research but it was a long kind of process
so okay
so this
is these are just a few slides from
a talk I
gave at Disney when I was there 15 years ago which is
all the different ways
new ways companies have invented to kill that
goose so one of them as
well let's just eat it forget
about those eggs or this
is a good one our latest innovation
is a goose that lays eggs of solid gold that's a distraction from our
core and we have no budget for goose related expenses
on that note we'll need the feathers and liver for another project
only one
gold egg every 12 or I
want gold coins this
is a Disney one I want gold coins rather than golden eggs
or I want platinum eggs
no you can
buy platon with the gold from these eggs
make the goose a manager
give the goose a deadline
require the goose to explain to
you how they're going to make the next egg
every one of
these this is as this is just at the level
ridiculousness that's going on they're missing the point
that nobody who does
these kind of worries has ever laid a golden egg
it's not their business to
deal what their business is is to count those
golden eggs after they get laid okay
so I'm going
to going to wind up
yeah I'm gonna wind up here
just give me an example of a
process that we use all the
time back in the 60s and the 70s and we still use today and
there enough I
always look at the reflection from the
heads in the room so
if there's a lot of reflection it means there's white hairs
and no hairs
and what
that means is there are a fair number of people who still remember who
ring Wayne Gretzky was right greatest
hockey player who ever lived and
yet a couple of neat
ones one was you miss 100% of
don't take asking
player in the world the only way like 106
the pounds he was tiny compared to the rest of these
but his best one was was the good hockey players
go to where the puck is and great hockey players
go to where the puck is going to be and he didn't
mean tracking the puck he
meant get to that place in the rink where somebody
can pass you the puck that you can shoot a goal he
was better at anybody at knowing where that place
would be and his teammates would feed him in bingo and
so the thirty year Wayne
Gretzky game is to have a glimmer of an idea take
it out thirty years where there
is no possibility of
incremental II
think worrying about how am I going to get from where I am now to this
idea right that is the the idea killer of all
time how is this incremental to the present
and the answer is forget it
don't worry about now the president
is the least interesting time to live in
so little glimmer
of an idea I had and the sentence would be it would
be ridiculous if we didn't have so
I had a little idea about children we
were thinking about personal computing in the 60s and I
started thinking about well what about a computer
for children personal computers are going
to be the next greatest invention after the printing press
we have to do something for children and you don't want
children to lean over a desk
we want them to be outside and etcetera
etcetera so I thought it'd
be nice to have a little tablet and in
this scenario they
are actually learning about orbital dynamics at
age 12 from having written a little
version of space war themselves and
the two computers are communicating by wireless
so that was a fairly you know just turned
out ARPA was fooling around not just with the internet
in those days but called the
ARPANET then but with wireless versions
of it and so
if you take it out 30 years
the puck is going to be there
thirty years and Moore's law
is going to go like that she been well
covered and
predicted in 1965 out to 1995
answers yeah goddamnit no question
1995 there is no way we are not going to have a
tablet computer no way it's just
happen we don't even have to worry about right now what
because what we have to do is to figure out what
it should be
once you start thinking about it then
the next interesting part of it is bring
back a more concrete version so
out there you can do pie-in-the-sky what about 10
or 15 years out
what can we do then and the answer is yeah we can do one
then what would that be like well we don't know
have all these problems including the user interface
problem no we've never had something like this for
the general public now here's the cool
thing about Moore's Law which still exists today and you're
seeing it in Hana and
that is you can buy your way into the future so
Hana could
have appeared 15 years earlier if people
have realized that
it isn't completely inevitable Hana
is Moore's law applied to discs
namely you don't need them and
we have to do all of this
software stuff now so we want
to get there earlier at higher expense because we're going
to save more money this is where the simplicity comes from pay
more it's like a fram oil filter commercial
by those five
dollar things frequently and you don't wind
up with the big expenses later on and so by
just spending money you can take something
that's going to be a couple of thousand bucks
ten to fifteen years you know out in the
late 80s you can bring it back into the 70s
for twenty thirty thousand dollars
so that's what we did that's where the
Xerox thing that looks like the Mac
prototype this laptop
because we couldn't build that
display in 1971
something that would do everything else so
that's where Mac type personal computers
came from and the we had a genius by
the name of Chuck
Thacker who
was able to actually build this machine
in a little over three months and I
another genius working for me by the name of Dan Ingalls who
could take some of my object-oriented ideas
interface ideas and actually make a system on
this prototype laptop
that happened to be a couple of cubic feet big
and I was kind of in there
doing this I
wrote the first interpreter for the first object-oriented
language as part of this thing
there and
we made 2,000
of these now here's the two things you get
by paying this money sounds like a lot I mean zahrk's
went batshit twenty-two thousand dollars
and you want to make two thousand of them
we said now this is nothing it's
nothing they have to be all networked together
and all this stuff but you can do two things everybody
has a supercomputer what
does that mean it means we can do zillions of experiments
without having to optimize we can do 10 15 20
user interface experiments today
so by the way these don't cut it
you've got all your people working on
things but these are the machines of the past so
you cannot do a new user interface on this
right because you haven't given them the super
computers you're trying to get
next stage user
interface out of last year's machines and
the other thing you can do is if you do optimize then
future apps and we made quite a few of them the most famous
one is Microsoft Word which was actually made in 1974
and that
very system was the one that ran in the
okay
so that's
how you get the the puck into the goal and
this particular process is
needed because before you can deal with the present you have to
with the future then you can bring back into the
present that blue plane version of the
future rather than trying to increment off the
off the pink plane
okay last slide give
you something to think about here is this idea
of thresholds so
how many people have seen curves that look like these
progress against
time right everywhere
reading scores
test scores people love these
yay
oh no yay
oh no
it's
bad because our nervous system is only set
up for relative change and
in fact there's cause for cheering
if that's the threshold but in fact for reading
threshold is this
this is all oh no
doesn't matter whether it goes up or not
because there are many many things
that where you have to get to the
version of the thing before you're doing it at all in
the 21st century it doesn't have help
to read just a little bit you have to be
fluent at it so this is a huge
problem and once you draw the threshold
in there immediately converts this
thing that looked wonderful into a huge qualitative gap
and the gap is widening and
we have two
concepts that are enemies of
what we need to do perfect and better
right so
better is a way of getting
fake success we had
improvement see
it all the time it's the ultimate quarterly
report we had improvements here and
perfect is tough
to get in this world so
both of those are really bad so
what you want is what's actually needed and
the exquisite skill here
which I'm going to use these two geniuses
Thakur and Engels to
it I'm going to call that the sweet spot the
way you make progress here is you pick the thing
that is just over that threshold that is
qualitatively better than
all the rest of the crap you can do you can spend billions
turning around and
once you do that you widen
up you give yourself a little blue plane
to operate in and for a while everything
you do in there is
something that is actually going
to be meaningful and will not
just bring lots of money I mean money you get automatically
out of doing this stuff even reasonably
well but the best thing you get out of this
stuff is a way of enabling people to think about
he situation that they're in better and not be overwhelmed