Transcript (view)

0:03
colegio background topic
0:07
this year’s proposition is that design matters more than that
0:13
it’s a pretty strong argument to be made for them asking for your merit mathis
0:16
proof given enough data and today there’s an abundance of data we didn’t
0:20
know and stewart brand once at the rate information in the replace just changes
0:25
your life
0:26
probably harness the power of data analysis modelling can succeed he’s
0:30
pretty definite makes revitalize education
0:33
abuse they can visualize undermine economies and still elections surely the
0:38
algorithms that they did a map that’s what the steeply leaks
0:42
but your life or changed by itself
0:45
bruce now defines design as the human capacity to plan and produce desired
0:51
outcomes
0:52
math informs the design compels with a design basket do it’s thing
0:56
poorly designed experiments collect the wrong data if the data is not listed
1:00
acted upon me as well not every bit have been coached in the first words so this
1:05
is the question we’re going towards a new beating to him
1:08
the proposition today
1:10
is that
1:12
i want to get this very very clear
1:14
design matters more than mass
1:16
so this point we need to do is look at the scientific board room
1:20
how many people here resisted and we don’t believe design matters more than
1:23
that
1:27
behind him
1:28
he dropped administrator allen
1:31
you get both
1:33
design matters more than that
1:39
them
1:46
we should do like how much they know forty-five
1:48
those of you coming in raise your hands if you think the site has been a map and
1:52
get your own benefit
1:54
this is exactly where the math
1:57
right well you guys are still hopes for triggers in your from here
2:02
forty five
2:04
raise your hands if you think now if you disagree the matters more than design
2:18
well as a lot more than over here they really didn’t know where they were with
2:20
the system
2:22
does this get at the two reflective reflecting character you know bathroom
2:26
i’m on the map site to uh…
2:29
that’s about things but
2:32
i got about forty
2:34
part did into its second i think it’s about forty five to forty
2:40
figure kids on the boat like i said this is the most profits for review of the
2:44
election should be obvious what the real problem with the israeli collection
2:48
advise more than that were designed
2:51
all right
2:54
opening up our last we have to jean georges the teams over here uh… water
2:58
really doctorate dot he is a senior research scientist linked into the back
3:02
of the carnegie mellon
3:03
i_b_m_ a_t_t_ doctor alexander grades were convicted of provinces is type of
3:08
nasa in nineteen eighty three
3:09
pretty good people for that argument
3:11
or left
3:12
julie steel is the editor beautiful visualization co-author of
3:16
designing data visualization she’s the curator of the data visualization
3:18
showcase
3:20
uh… should be designed track that she hosts yourselves with chairs tracks
3:24
unclear story it is
3:25
dr assorted douglas panel and formulate user experience for google analytics in
3:30
a variety of other products
3:31
both ably equipped to talk about
3:36
so the first person in the design team which tells of four minutes or less
3:39
by design that is more the mass sure so i’m going to start by uh…
3:45
so designed matters more and math i think they will go on the conference
3:50
it’s it’s pretty obvious for investing very heavily in mapper technology
3:55
and that’s great
3:56
and i sort of like if you add to my talk by yesterday
4:00
when i came out and we will look around the big bear market
4:03
uh… it sort of felt like a old cellphones uh… if i could use images
4:07
imagine those big cell phones
4:09
uh… was big bird cellphones ahead very promising technology that we know and
4:13
sunday change the world
4:15
but i think was pretty obvious in people reason that the vice the knew there was
4:18
something more to this like something else had to change it really wasn’t
4:21
fitting in the analyze of people using it
4:24
the same way i think that’s where big date is now properly sell a lot of
4:27
people in this room in this conference in silicon valley in general
4:31
are selling gary hart technical problems
4:35
i think if you look at the cellphone or another example use the home thermostat
4:39
so some of us in my home i had these very novels those digital thermostat so
4:43
there’s all sorts of inputs in there
4:45
we have these things are lies where we can tell technology leads
4:49
so how do i think it’s produced
4:50
it’s not so i think design
4:52
really takes hold of a product into a market where there could be mass
4:56
adoption of their product
4:58
now that we all realize that this the technology were inventing now has the
5:01
potential to change the world radically so if you look back at the cell phone or
5:04
even better that little um
5:06
home thermostat devices sad thing isn’t necessarily at the interface socks
5:10
it’s really that there’s technology in that device that could
5:13
now listen hold homeowners lots of money but imagine if everybody in the country
5:18
across the world were effectively able to use that the vice that actually was
5:21
usable that the used
5:23
more than three percent of functionality
5:26
they could literally change how
5:28
uh… how much energy is needed to arm fifty homes which had to have a radical
5:31
impact energy consumption
5:33
so the same way if we wanted to get a technology to exist beyond this room be
5:37
honest conference beyond sort in bali
5:40
i think he really have to have a firm understanding
5:43
of all the necessary inputs to to make these things work in people’s lives and
5:47
that’s where design comes in
5:50
one other thing to note is the we have to imagine design is more than just
5:54
so what happens on the surface i think sometimes we think design is just about
5:57
how pretty addresses or what color or you’ll color of his elation uses our
6:02
fifth interfaces rounded buttons
6:04
uh… designed as a lot more than just that
6:06
um… infact if you imagine started designing a home so if you built a home
6:10
recently it’s like saying you know application design
6:13
uh…
6:14
if you think of p_c_’s design is just about color and how a graph looks just
6:18
like seeing designing a home which is about picking up the cobra carpet or
6:21
what color your walls you you wonder what hardware using your kitchen
6:24
but really if you’re going to design a home well you have to understand the
6:26
needs of the families who live there
6:28
you know do you have kids are gonna have more kinsey have now how do you want the
6:31
rooms configured wears a light coming in
6:34
and that’s where designed to help big data shine which is understanding who’s
6:37
using the product where they’re using it
6:39
where there is uh…
6:40
silos with an organization that could uh… use data
6:43
almond so if you really want a locked the power of big data you have to
6:46
utilize design
6:49
nicely done with that
6:51
minutes fears
6:53
uh… who views canisters
6:56
all right
7:00
uh…
7:02
at the wilmington
7:05
and the like math
7:06
so
7:08
and state
7:10
right blair
7:12
with a bit
7:13
also aircraft oil rusty
7:16
as they have professor d
7:19
and efficient itself
7:21
what that well mesmo
7:25
mask and the magic
7:26
uh…
7:27
that you basically can do in a purely human way
7:30
one of the uh… things and comprar two thousand really
7:35
we’re looking at a problem in astrophysics needed users cystic
7:39
turns out to be critical mass of physics
7:43
tend to the nine times faster
7:45
faster out with them
7:46
there’s no way you can do that with anything humans
7:49
unless he gets into the ninety mins annual maintenance to appease the
7:53
computation something
7:55
stomped
7:56
google is magic
7:57
who got all
7:58
happens instantaneously
8:01
you can search for anything that speed rank that’s very indices these are
8:06
are part of the math of the fish in alan’s
8:10
and how
8:11
new kinds of ways to represent the date of that
8:13
are explained by method i didn’t values its
8:17
uh… their questions you’ve
8:18
they cannot even deposed without matthew can even asked the question properly
8:24
except in a mathematical way
8:26
if you’re trying to diagnose
8:29
cancer
8:30
machine learning for example from uh… numbers obtained from people’s
8:35
uh… blood samples
8:38
you can’t predict something how do you
8:40
even talk about contacted me you can predict something
8:44
need the catalyst of probability
8:46
all you need to understand the concept of risk you need
8:50
also confidence how can we be confident
8:53
don’t want to tell someone may have cancer when they don’t mice for self
8:57
you need a national conference
8:58
that is a very delicate thing that uh… decades
9:01
teason
9:02
pose that quest
9:06
and finally uh…
9:08
the iranians in the context of dana l
9:10
the i_r_a_ no design helps humans do things better but
9:14
there’s certain things that newman’s just can’t do very lively
9:18
wheatley i can only see certain things
9:21
and his these lean is misled
9:24
so you cannot see into more than two or three dimensions
9:28
for this accident if it’s twelve dimensions or a ten thousand dimensions
9:32
u
9:33
need to turn to math
9:35
to even asked certain questions
9:37
uh… and finally
9:39
even when you can you think you can see
9:41
on your data into a three dimensions
9:44
their patterns that you’re gonna see that on accident because your body is to
9:48
will change this e patterns
9:50
and and will be misled
9:55
right with uh…
9:56
despair
10:00
how do you feel this is i know d
10:03
and went out
10:03
than everything that math doesn’t matter recently saying the design matters more
10:14
you send the knife can do magic but you can’t even if you know and i would agree
10:18
with that but
10:19
what design does is take that from something in human back to something
10:23
came in and that’s critical because it’s humans
10:26
that have to make decisions on how to take action based on information because
10:29
of the end of the day that’s what data is its information and and what use is a
10:34
if you can then change your behavior
10:36
uh… are just based on that information
10:39
and who’s doing that it’s humans most of the time sometimes it is machines
10:42
typically guided by humans are built by humans
10:45
at the end of the day information as a human thing so
10:49
not concerned that i think he missed can deal with the end of day designers the
10:53
language that translates that back in the something actionable people
10:57
uh… and that’s why it matters more
11:01
out you also mentioned that certain things humans don’t do reliably the eye
11:06
can see certain things as easily misled well yes exactly precisely that’s why we
11:10
have to decide to take that into consideration
11:13
we also can’t john
11:15
twenty feet in the air that’s why we build stairs
11:17
design as the mental status
11:20
that lets us get where we need to go
11:25
wertheimer u
11:28
should
11:29
action
11:33
so yeh questions that they can’t be posed about math i’ll get back to you
11:36
know we need certain mathematical concepts again that matters
11:40
if you’ve read the new york times lately or
11:43
and he had a kind of journalism that within talk about you know
11:47
thirty percent everest fit for this and then one-in-eight for that
11:52
doing this comparisons in talks to positions
11:55
that’s not just not because they are two percent of one and one eighth of the
11:58
other is math but it doesn’t make any sense because you can actually compare
12:01
those apples two oranges mines
12:04
so choosing how to frame something that’s a design issue
12:10
design matters more knock them out as a matter just the design matters more
12:15
uh… so i would make an analogy here and say that um…
12:20
if visualization is this sort of chopper the transportation the easing
12:24
ideas are the payload bay information is the thing and albicans unwrapped they’re
12:28
useful for transporting ideas but they don’t communicate our movement from one
12:32
bringing to the other way out some of the anti-gay information from one place
12:35
to another place and design is that that mode of transportation that that makes
12:40
things human
12:42
and and that’s when i saw a stop there
12:45
yield the balance right
12:47
well yielded all rights monica
12:50
like three to be here defending that’s on there
12:53
personnel
12:54
they are scientists my backpack me
12:59
you might like to ask me is this insecure about mine
13:04
uh… sellout it makes sense that i was going to be on the second half
13:09
and after having that discussion without there
13:11
uh… i’ve been thinking about it and i noticed something i noticed that
13:15
outstanding a lot of time at work
13:18
popping today dot find this
13:20
telling them to visualize debate now
13:22
to uh… uh… clean up their charts and graphs making simple and easier to
13:27
explaining to digest
13:28
in uh… and to make sure that would be a product that they weren’t like
13:33
recommender systems make sure that well designed to they speak to that use
13:37
so that they mean by strengthening many wonder well
13:40
realistic
13:40
design pattern oregon
13:43
so un
13:44
i thought about it and then you dont mind
13:48
completely cds on
13:50
because this substance is already there
13:55
scientists and engineers spent a lot of time months living here’s
14:01
putting together they are packed or a billion dollars an hour
14:05
visualization
14:06
and the background of that experiences the fact that i think he’s location
14:10
so up by the time we got around to it this substance was already there
14:15
sellout dot substance is back
14:18
uh… it’s the other than that
14:20
howard those information systems
14:22
it’s uh… what makes at six in the alley
14:26
uh… if those uh… gardena robot space travel right that’s why we live in the
14:31
future
14:32
uh… an
14:33
that’s webkit this time
14:36
accidents and that’s how we can at the site leads by campaign
14:50
expelling
14:50
wakes me stuart
14:53
to understand that universe around x_-rays it’s how we communicate approx
14:59
time across languages across pollsters
15:03
it’s matheson’s sent stocks this is what matters more
15:11
subduing arguments so uh… it’s time for both starting with you sir
15:15
uh…
15:16
yes all we will start
15:17
where uh…
15:18
monica left off for math is substance nigeria’s substance
15:22
but i would say those sort of
15:23
backs original point if we want that substance to live in this world in a
15:26
meaningful way
15:27
beyond
15:28
kill the brilliance of of the folks over on the side of the room
15:33
normal people or people who don’t get that they don’t really want to see that
15:37
substance they want to see that substance invented in something
15:40
that works in our lives
15:41
so if we want people who don’t understanding data or don’t other
15:45
understands
15:46
nearly bands are rhythms world venting everyday which are very important that
15:49
we want them into actually utilize that
15:51
that’s when we have to surround those things
15:54
with was the design friends not just the interfaces understanding where those fit
15:58
in the lives of those people where we can put those on your rhythms to make
16:02
them more successful new job
16:03
whether that’s predicting how customers are gonna buy the product
16:07
or whether it’s predicting when the next operate double fu is going to be or
16:11
or something like that you have to really in bad that in a meaningful
16:14
design application for people to use them
16:17
unicorn i want to say to you is uh…
16:19
no particular now since being so i did a lot of work actually monoliths
16:23
around this and i would say
16:25
you know if you just took up your predictive analysis and that has all
16:28
these confident company intervals
16:31
that was one of the hardest thing eight designs how defect that we communicate
16:33
that
16:34
two people
16:35
they think in the same way bad design
16:38
could manipulate people
16:40
i think maps to do that as well so i’d accumulated work really hard to get a
16:44
design which i could effectively communicate the people that the data
16:46
you’re seeing there is some intervals confidence in here you have to
16:49
communicate that really well folks so
16:51
designer really is important on top of math to make
16:54
the math the substance were making meaningful relevant people’s lives
17:00
that’s it
17:04
bonaventure
17:06
the study few seconds left in standard amaze you said you and that’s why
17:10
resigned about landscape
17:12
usually people are like fresh each other and isn’t that right
17:15
all right at the starbucks with it
17:18
poor uh…
17:20
sure design is important on top of math
17:23
but the question is
17:25
which is more important cell
17:28
group will search with your ad there have grill
17:31
without great interface beautiful interface simple
17:35
but they’re uh… sixteen months behind it didn’t reach you can see it’s
17:40
or we’d still like are
17:42
called fall pagerank settles the computers
17:46
maximising marrying
17:48
predicting that i’ll based on your keywords what should be shown to you and
17:52
so on
17:53
and she faces kankakee
17:59
dot org
18:01
safety
18:02
there was a you know
18:03
yield the floor
18:06
rights june
18:08
uh… no
18:10
after the start of math is universal as how we communicate across time and
18:14
cultures
18:15
i call what’s the word that i can use on tape
18:18
and
18:21
it’s not universal their cultures that have words for one
18:24
and to and then more black at baruch derek weeks of turning in different
18:30
cultures that that are certainly not universal uh… design however cognitive
18:35
visual perception
18:36
contains principles that are universal we know this then to soften samples of
18:41
grouping uh…
18:42
and ranking size ranking things like that
18:45
those are universal across cultures
18:47
and tribes
18:49
time
18:50
uh… so i would say to them i think they’re gonna try to hard-line what’s
18:53
more primitive
18:54
design wins every day
18:56
uh…
18:57
i would also say alex taylor fight
19:01
it
19:02
i read a avenue
19:05
alex your point about the most serious and google the great inner cities or a
19:08
strong search engine with a copy of this
19:10
and if you’ve seen those being television commercials lately where they
19:14
itil clients have sinned you know people say which side of search results do you
19:18
like better and most of them she’s being because it’s a big commercial
19:22
but wire more people still using google
19:25
because the interface makes it worthwhile that design matters
19:28
and people are willing to sacrifice a certain amount of math error accuracy
19:33
when the interface makes them more comfortable as more usable
19:37
and now i have thirty seconds left panel and then ok nineties authored measles
19:41
sousa so being a former employees of google they also know that
19:45
they put huge investment in what seems to be a simple interface
19:50
uh… into the design of that so simplicity is it
19:54
no design
19:55
in fact
19:56
you could
19:57
characterized in the best designers in the world
19:59
really embrace simplicity
20:02
fact deterrence is probably the most one of the most iconic designers in the
20:05
nineteen fifties
20:06
isn’t going to see today threw up on the ipod he said you’re not going to design
20:10
into all the elements that are necessary taking out of that so simplicity is
20:14
definitely something that uh…
20:16
and that’s why design
20:17
there’s more but i will not vote tells me it’s time for
20:20
another about monica
20:26
left that we’re seeing it deserves dot is con artists compensation is it
20:31
possible which week
20:35
uh… cell
20:36
there’s a friends though that i don’t agree with with since that time designed
20:41
has a monopoly over beautiful effective communication
20:45
because if you think about it at exactly what is
20:49
you all effective keys
20:53
we have allen’s we have simplicity we have dvd bank does not constitute some
20:58
cheap
21:00
and uh… i have a what’s on my side
21:02
uh… said that matches the poetry uh… logical ideas
21:07
so com
21:08
and to bring up the story about that stopped i actually wanted to get the
21:12
estimates that wednesday well-designed high i’d love it com i’d like to get it
21:16
i think that the reviews that said that now
21:18
you know i’m not
21:21
so behind
21:22
hyderabad
21:24
and i said well i you know why i lobbied for design but i’m not completely over
21:28
it
21:29
right it’s not worth freezing
21:31
so uh…
21:32
but you know they’re going to get a right they know what happens
21:35
may they don’t like it was hard at all
21:37
it’s about learning to us that they know that the court
21:40
so they’re gonna get it right then i’m going to buy it
21:43
that goes to the other wider adoption of a product yes you need a map
21:48
so uh…
21:50
is not about the distinctive the state st
21:55
it’s about
21:57
it’s about making sure that their website
22:00
actually works
22:01
i bet you do with the leadership that designs for any bears they believe that
22:05
we don’t ship marks
22:07
right com it’s about substance it’s about
22:16
faster balls some designer somewhere is going to go read a book called fifty
22:19
shades of blue rate now
22:23
acts we have a few questions so you’ve sent me some i have a couple years uh…
22:27
first
22:27
also had a cough retailer says this debate needs better definitions that’s
22:31
the point of debates but he says
22:34
we should’ve called this algorithms versus pretty this i think i know which
22:37
side he’s gonna vote for up
22:39
so given that’s the case uh… how would your arguments change design team
22:44
if the title of this was
22:46
algorithms versus pretty dance
22:49
so i think that
22:51
unfairly pigeon holes design
22:53
uh… it’s sort of
22:58
i don’t think that’s accurate to do bomb
23:04
i had ended about how beautiful visualization and one of the things that
23:07
i’ve done a lot of thought into when does that was coming together is what we
23:10
mean by beautiful and i’m going to go ahead take delivery of uh…
23:13
substituting beautiful victory message i’m not sure as you know where
23:17
and
23:20
beautiful
23:21
as we define it for the purposes of that book is an alligator solution of some
23:24
kind of problem right that’s
23:26
that’s the eighties is sort of an elegant and simple solution and that’s
23:30
what design really is so
23:32
i think
23:33
they tend to brand design and simply treating us with uh…
23:37
was that three transparent tentative reduce its respects
23:41
and and i think that’s unfair and span incomplete inaccurate inaccurate
23:48
stupid uh… yours
23:52
from sovereigns versus proteus culture fights
23:55
of the better separately
23:58
seetickets
23:58
unfair characterization to say that design is
24:02
just produce yeah i think that’s a big
24:05
ac standpoint
24:12
because it’s not
24:16
let me ask you this question
24:20
howell does design
24:22
influence the formulation of mathematical algorithms
24:25
taking the other side here
24:26
we always look for reducing something elegant region for example
24:30
is that just decide at work in math and its own does that mean that designed is
24:35
an inherent part of the moment
24:40
moon
24:40
uh… actually true
24:42
uh… elegance is uh…
24:45
is would we all know is our programmers
24:48
uh… actually has functional
24:50
value
24:51
amid leaves few fewer bugs it leads to greater understanding of code
24:56
in the end greater longevity of code
25:01
it is true that
25:02
programmers and mathematicians are
25:05
did the best once the great ones are designers in the language
25:10
uh… whatever computer language of the language of maths
25:14
so yes absolutely uh…
25:18
uh…
25:19
yeah so
25:20
he did it
25:22
uh…
25:23
we’d like to simplify the forecast because we want the better model about a
25:26
mathematical model that explains the world a around standouts
25:30
that’s why com we’d like to simplify it
25:33
small beautiful point
25:35
because it’s more power
25:36
but it’s difference is that
25:40
dated
25:42
reducing
25:44
the difference is that it is grounded in something quantitative in the internet
25:48
allows you to know
25:50
what you’re doing better or not
25:52
on your intuition
25:54
if you’re really smart is great but even when you really smart may not be
25:58
uh…
25:59
may not be right
26:01
re we aren’t needed in our kind of designed by
26:05
being able to quantify whether something is better
26:08
design guys
26:10
but the idea that elegance in math is a former designed are looking for a
26:15
against chisholm
26:16
anesthetic
26:19
as they are reaching some kind of downtown uh… ad because i i would
26:23
agree that
26:24
it’s certainly possible to retaliate sitemap
26:26
uh… and that that is until i love that idea that the best mathematicians are
26:31
designers i would agree with that
26:33
and and honestly the best designers are mathematicians we have certain
26:37
principles like the golden ratio where it you know
26:39
the pesticide doesn’t blame apso
26:42
out off or imagine these two blacktown
26:46
is were being forced to choose between one and the other or to rank one above
26:50
the other
26:50
i still have to go to sign every time
26:53
and eyelid
26:55
talk about a condition for about models in trying to get to a simple elegant
26:58
model uh…
27:00
and i would say
27:01
we’ve got this interesting problem right now i a specially enterprise
27:05
we have a lot of models that are more accurate or at
27:09
resulting better vindictive results
27:11
but we don’t always know why the workfare not transparent for the black
27:14
box models and a lot of times companies will choose
27:18
the last predictive
27:19
but more transparent model opted out of litigation fears because they need to be
27:23
able to stand up in court and defend why they have
27:26
done something in a certain way
27:28
so when it comes to the east you know we sometimes have math
27:32
that is better in the sense that it’s more accurate but we can’t explain it we
27:36
understand that and
27:37
as we actually to use the less effective map
27:40
because we we can see it and design a rabbit uh… and that’s me says a lot
27:48
uh… not protect bugs
27:51
so questions so that’s why i was growing up it indicates a new
27:53
riza reinanto something nasty
27:56
right then you
28:23
so the reader that’s of basically isn’t really about measure ability and if so
28:28
that is designed as the quantifiable mathematical process does not
28:32
then that’s a something equally interesting that
28:35
a little of the matter is the first
28:41
you lower
28:42
get a p_h_d_ engaged see i human-computer
28:45
interaction
28:47
use more of what you are news you gotta do user study
28:51
you’ve got to quantify
28:52
where they’re your idea about whether it’s better not
28:56
stroup c
28:58
was a good idea so
29:00
actually turns out the sign is a subset
29:10
europe
29:14
yet
29:15
society
29:17
it is about the uh… you can sometimes you you can’t you can look at
29:23
uh… what uh…
29:25
happens when you have decide without that
29:28
you get
29:30
highlights
29:41
catalysed beautiful eighty
29:43
acceptance that it would be fun
29:46
it
29:47
beds right now
29:49
n an
29:51
it’s it’s installing skilled sings he’s like
29:56
how except that i think what work
29:59
not not yet
30:01
uh… hanging
30:02
they well eventually wants to get them at night
30:06
document
30:11
it’s
30:11
that was uh…
30:12
go ahead trent who had sex with him
30:17
so
30:19
anybody’s reviewed
30:23
ideal for working together design right
30:26
maybe just cos emerging all-time that are trying to address that problem
30:31
you know there’s all sorts technology i think were
30:33
we’re developing here that
30:35
in the same way that you know of pretty now need designer pretty mad enough to
30:40
work for a great technology technology needs designed to work just as much as
30:44
you know around
30:46
the other thing to or
30:50
is that we need more time at other instances train as a few minutes ago and
30:54
i
30:55
staunchly defended design is the more than just the specs
30:59
and ask them by that design is more than two suspects that encompasses a specs
31:03
and please try to qualify for me the mona lisa
31:06
and tell me how you doing track five five some of these things are just so
31:09
moving
31:10
and service or all
31:12
and designing compasses those things and you can’t qualify and said no design is
31:16
not a subset
31:20
so i’m asking requested your sir
31:23
uh… but i i thought you can things that are for the point in the audience
31:25
of uh…
31:27
let’s talk about local mocks so you guys are these are the cause of the local
31:30
maxima i can take something and i can optimize it until its local matson
31:35
but it generally takes i’ll be giving tuition or a art change
31:38
to move from a local cable net
31:40
obliterate he was a really good inspiration she’s a really adopted
31:45
and i would asking
31:47
is it more important to optimize to a local maxima or to have an inspiration
31:51
takes you to a new global maximum or more things are possible
31:56
you guessed
32:00
can you repeat the question
32:05
so three-year-old local taxes
32:08
imagine you’re on a hill
32:09
and the water it’s down the hill too late on the side of the hill that’s the
32:12
lowest place the water could run to
32:14
but it could also run much lower down the hill to the ocean
32:17
so that the lakeside they feel suddenly get the local max of west side of the
32:21
hill
32:22
it’s the lowest place in that area
32:24
computer algorithms tend to be really really good at solving for a local
32:28
maxima so their webpage designing the whole budget multivariate analysis of
32:32
testing optimize clicks and the algorithms awesome attached but if i
32:36
wanna go from eight percent conversion a twenty percent conversion big
32:39
achievement business model
32:41
and that’s the thing that the local optimization won’t do so generally from
32:44
analytics point-of-view uh… webanalytics walks look at
32:48
optimizations it as a machine problem
32:50
that innovation is a human
32:53
yes i did tell a kid back on a plane to several
32:56
and historical moments of science
32:58
uh…
32:59
is also millions is iran i hate to see you design is the human element and
33:04
ultimately as humans but we do is tell stories
33:06
and if you really interesting points in the history of science where
33:09
the math couldn t
33:11
evolve until we change the story
33:13
so one of those was haven’t
33:16
the
33:17
the solar system right whether that there is a lot about the scientist
33:20
except for a long time but i don’t think we have some of you know it’s actually
33:24
son revolved around us
33:25
and they were trying to figure out or red cedar tommy has written on this for
33:28
a while
33:29
and you know of course they were also seeming perfectly circular orbit
33:33
it wasn’t until we change the story
33:35
assembly arena actually were not the center of the universe may be that we
33:39
are on the side
33:40
that they were able to start doing different kinds of complications on i’m
33:44
planetary orbits and it’s also been allowed to get away from from perfectly
33:48
round orbits insult the elliptical and so we now know are close to the church
33:52
but it wasn’t until we shifted that story in our minds that that the math
33:55
evolves and another play i would point here is the germ theory rate
34:00
until we we change the story from what we’ve got these humor is weekend applied
34:05
leeches in
34:06
so that the other something so small that we could see it that was
34:10
transmitting disease
34:12
only then good science and math behind it really get to where it needed to be
34:16
so
34:16
the storytelling and the design that allows us to think about that are are
34:20
the critical factor the lead after a four e
34:23
you’re listening to it um… monica i think uh… reference the fifty shades
34:27
of blue which i think she’s pulling from the blog posted double monroe when he
34:31
left nipple thought that was it or not
34:33
writes about bombings you know is the first real designer baby lol and he left
34:37
sorta publicly forestry wrote a blog post about how as a designer u_s_
34:41
forestry because there will be sweet at this
34:43
approach where they
34:44
they were designing through optimization
34:46
especially on search
34:49
so double apparatus blog post in biblical teaching within google and
34:53
overtime who will recognize that there is so much you can do with optimization
34:56
it’s valuable to the but you also need to bring in this sort of human element
34:59
redo take a leap of from intuition
35:01
north until then in some other things happening within the competitive
35:05
landscape that google really understood that it needs a balance of both firms
35:08
optimization me perhaps the amassed way of
35:11
designing things
35:13
uh… versus sort of the human more intuition model
35:16
and i i think do election shows an increase in design with that
35:22
choose
35:23
localnet
35:25
what that means start not necessarily
35:28
might not design doesn’t have a monopoly of a quantum leap sickening about
35:32
hearings or on
35:34
mathematical calculations that people realize were too weak for the physical
35:39
caught up with them
35:40
realize that you have a tax return expanded your signal like with that
35:45
infecting yuri learn a lot of people physical up
35:49
expectancy in convincing pain
35:51
so design doesn’t have this monopoly over beats about also can offer the
35:56
quantum leaps and as far as getting stuck in the local uh… minimum
36:02
outlets offer
36:04
solution to that
36:06
has a solution to that right there’s a bunch of things you can do to avoid
36:09
getting stuck an example
36:13
uh… there’s there’s seemingly damning there’s a lot of ways you can used to
36:18
and get all of that local mail
36:20
and get on moving to discover the best possible since
36:24
uh… and then as i mentioned it’s not about eighty testing it’s about the
36:27
contents behind it it’s about making sure
36:30
whatever you can see capturing works
36:36
local next are only a problem if you’re a function as uh…
36:39
not contracts strongly convex institutions so complex
36:46
at that
36:48
norquist’s laura
36:50
questionnaire outside presided you first
37:37
unless europe wester
37:40
i’ll give you guys thirty seconds each
37:42
to to talk about whether the universe is designed or word code
37:51
okay i’ll let you talk to thirty seconds
37:53
about if mathematics is the language of design how’s that
37:59
this first
38:01
julia securities uh… anxious asses yes eyes
38:09
asking whether mathematics is the language of design a few minutes ago
38:12
arrested designed as a subset that hasn’t happened subset of design so out
38:16
also yasser
38:22
view a very good point was not ready to be each other
38:26
so i’ll tell more questions one is
38:29
you guys designed as we talked about if you rented experiments
38:33
did you collect crap data
38:35
because your design was really bad
38:38
ventura masses meeting statistics please firms
38:43
underscored
38:46
cynical on time
38:48
if you collected at the exit dot desai kirchner options
38:53
experimental design matters as much as user emphasizing the display something’s
38:58
so
38:59
but the design of the did an extra sexy you open to the court that ab about
39:03
information starters at uh… double-talk
39:06
debts and it’s about getting the right to use the information to the right
39:09
things at appetites prime
39:11
and so yet
39:13
it’s nice to have information as i saturday as i said the other cuts and
39:17
apply algorithms to that but if it’s corrected poorly access i collected
39:21
poorly
39:22
stored poorly
39:23
inaccessible at the time that you need it
39:25
and its use
39:32
burst
39:36
is here
39:36
you designed here experiment poorly
39:39
it’s because indian redone math book
39:49
there is a very attentive and expands
39:52
question yes sir
39:53
if bottles nastiness
40:13
so mathematician or design of it for cross river and swims and the
40:17
mathematicians those of the month average the rivers three feet deep sugar
40:20
land general
40:26
passenger although i have to miss it
40:29
therefore designed by them
40:33
france now
40:36
antibodies produced under the word word would design united flight jacket and
40:40
her clear
40:42
float on a few ferry boats what third-party no schatzi
40:49
hit acting up
40:53
his
40:58
sites we have that’s not the speaker
41:01
uh… rebuttal
41:02
walls to each and we started designs of the math gets last word because we begin
41:06
the design
41:07
which it would like to go first
41:12
yesterday i think this is outside the home sick but still
41:15
so our mistake the one example of nasty arm
41:18
so i know nest pretty well google ventures time salam my for friends and
41:22
colleagues are on the google venture team
41:24
and they work to really closely with ness
41:27
and ne spent a lot of time found what you might think ah…
41:32
talking to users ready to talk to people to today really figured out how
41:36
people wanted to use
41:38
nasty within their home
41:40
and i think now sort of got the fact that design is better than math because
41:44
they knew that they can give in to people that this device was something
41:47
that you know they really wanted
41:49
they convinced through in the ease of use to design in it worked good enough
41:53
that they always knew they can add more math on top
41:57
and so i think it’s a good example of the lease from a market point of view
42:00
they understood that
42:01
people
42:03
are looking for things that they want to be crave and nasa provided that device
42:07
for for these folks
42:11
outside
42:12
blizzards
42:13
yes
42:15
so uh…
42:16
i’m sorry to have to deliver the final crushing blow but
42:20
big date up
42:22
it’s not there without decision almonds
42:25
in fact data can even be talked about properly without
42:28
problems this takes
42:31
in fact there no computers
42:33
without
42:34
math the mathematics of
42:37
computation
42:39
and programming languages artefacts math
42:44
the data or two
42:45
have as it’s only basis design
42:49
that’s the public library an aspect of boats that are
42:53
brutally design
42:55
the alternative is that beautiful recent so we can see it
43:00
but lately
43:02
julie classroom
43:04
graduate at patronizing and i said
43:06
jet is at hand
43:08
you know well here in the context of the data and we’re all here in the context
43:12
of putting kids were car companies and for ourselves
43:15
and
43:16
and back wires that the excess holds a lot of people people have to make
43:19
decisions people have to take actions
43:21
and the thing that makes it accessible
43:24
uh… once you apply or not because again
43:27
nothing the map doesn’t matter it’s critical
43:29
but as i was more critical because once you have that information once you apply
43:33
those algorithms you have to be able to take action on it and it’s timely
43:36
fashion at the right place at the right time
43:39
and design is the
43:41
is everything that allows you to do that and do that well
43:48
so i started without books online thing about thing about now
43:52
beat la guardia elected output dvd’s
43:59
you might think experts
44:02
we all use matthew fifty-eight
44:04
to predict what
44:06
to handle any
44:08
is more than forecast like these
44:10
it’s likely expression
44:13
is using your mind
44:15
to solve the biggest mysteries we now
44:18
so that statement right but uh… i my crack addict so uh… here’s how i like
44:22
to think about it
44:24
an iphone without design
44:27
isn’t laughing
44:31
without map is too beautiful pink pants with a strange
44:37
cannot attack
44:41
fahey will prove to be close to a little voting so
44:46
ab first of all if you walk in after we started and you didn’t get to reassure
44:50
and what’s already please abstained from voting you’re welcome to laugh at the
44:54
good parts pluto vote twice
44:57
or don’t vote if you look for now raise your hands if you believe after all this
45:01
debate
45:02
the design matters more the math leads them up design matters more than that
45:17
as i think it’s up to out
45:20
gente about fifty day raise your hands you think that map out of more than
45:23
design
45:25
outside stuff
45:27
it’s look at what’s the the first time
45:29
i think i mean togethers marginally
45:32
since i’m counting for the massive
45:39
thank you
45:40
graphic designer better portraits are tried to do a better but i’d like to
45:44
play
45:44
are for participants for trying to get a lot of your rates brick walls and going
45:48
at it
45:49
uh… i don’t think i think that uh… as as and i can be biased
45:53
particular blows the upper mass
45:56
because it’s
45:57
so
45:58
well clearly make so much for joining us
46:01
brought back to the main stage we have the fight
46:03
experts here