Tip:
Highlight text to annotate it
X
{\rtf1\ansi\ansicpg1252\cocoartf1038\cocoasubrtf360 {\fonttbl\f0\fswiss\fcharset0 ArialMT;}
{\colortbl;\red255\green255\blue255;\red0\green0\blue0;} {\info
{\author Balderas}}\margl1440\margr1440\vieww10820\viewh12260\viewkind0 \deftab720
\pard\pardeftab720\ql\qnatural\pardirnatural
\f0\fs22 \cf2 Part Two\ \
\pard\pardeftab720\ql\qnatural\pardirnatural
\b \cf2 >> ROSEDALE: \b0 A whole ton of people. Now when you sort
of pull the camera back, I am gonna talk over this video sound, what does that look like?\
\ Over the next ten months we spent about two
hundred thousand dollars. The total labor pool in our work list, which is the system
we use to build this, is about close to three hundred people all over the world and about
seventeen hundred actual commits.\ \
Every little, every little ball of light there is basically, you can see the mix of who the
manager, me in the case of that example I gave you, the developer and the person who
was doing the review on it basically doing all that work simultaneously, all over the
world.\ \
So there\'92s a classic, there's a guy who got the Nobel Prize in physics, he said only
at the acceptance speech more is different and I think that's an important thought about
sort of crowd sourcing this stuff. \ The point I\'92m making is that breaking software
development in tiny little steps, hundred, hundred-fifty dollars at a time, profoundly
changes the nature of the experience and this is the kind of tip of the iceberg stuff that
I think is so important to be thinking about as this group.\
\ But basically just imagine how weird this
was. That we've that we've built the software entirely using this distributed crowd of people
all around the world. \ \
So at the end of the day, \ \
[pause]\ \
There we go, at the end of the day as I said we've spent about two hundred thousand dollars.
Now Silicon Valley full-time equivalents, that\'92s like less than two people right?
It\'92s tiny fractions of a person. That's really cheap, to build a piece of software
like this. \ Now there was myself and my two co-founders
doing some of the design and management here but much less than you might think a lot of
it was actually done by some of the people we were outsourcing this to.\
\ [Pause]\
\ There's a lot of things that we've learned
and I want to now kind of talk about having built a complicated, now launched project,
like this. What have we learned from it, because that's what I think is salient to this audience
\ \
Well the first thing is that what constitutes good in a developer and I think in any worker,
ourselves included, is a little different than you might think and, and as we've gone
forward we\'92ve really started to learn a lot about this.\
\ So one interesting question and I love talking
about this, and I do all the time, my background is physics and software engineering as well,
is really like which person do you want a higher right? \
You\'92d think when you're doing, you\'92d think when you're building software that you
want to build, higher the pedigreed person. Certainly that man on the left has got a lot
of pedigree.\ But the reality, what we found out as we got
into, this project and we started working with hundreds of different people is that
that isn't true.\ \
[pause]\ \
The guy on the right, who has been programming all his life, he is in I think Bulgaria, he
actually is one of our developers, is tremendously passionate about and engaged in the idea of
working with such an interesting group people in Silicon Valley on a project like this.\
\ The guy on the left has already been pretty
successful and, the same can be said of just about anybody that's, been participant in
a startup and is working here in Silicon Valley. The guy on the left, emblematically, is a
person who has a high degree of skill but are they really going to use that skill in
the fastest best way for you in building your startup.\
And I think that that the answers to that are far more complicated than we believe.
In fact what if I further told you, just through transparency which we do, that I can just
drill down on this guy Krum and see that he's actually done seventeen jobs for us and he\'92s
made eleven hundred dollars. Am I starting to sway you away from Larry now?\
\ I believe in general that transparency is
a very powerful idea. We've really learned a lot about this. \
This is a funny idea but imagine if you are at a restaurant and you put up a floater above
everybody's head, in some augmented reality that showed the actual history they had of
tipping, how much they paid and how much they got. Think about how that would influence
the labor market if you will of people walking into a restaurant. \
\ [pause]\
\ Um... that's the kind of transparency that
we've created in doing this and again it has profound impact on the nature of how people
work together.\ In short if you expose that information to
everybody the market behavior that you get improves the quality of the product. \
\ [pause]\
\ And I wish we could do this in restaurants\
\ [pause]\
\ You also learn a number of other things.\
\ This is, this is another crazy thing you can
do with, with Work List which is the, Worklist.net is the product that we built before we started
Coffee and Power, and that we built Coffee and Power with.\
\ You can drill down on one person, like this
guy Hans. These are all real people they are up on our site; you can go see all this stuff
yourself so I\'92m not betraying any information about them because we do every day. \
Hans is a great guy who did a ton of work for us but you'll notice a characteristic
curve in his participation over a couple years. Peaked up at the beginning and then it sort
of sloped off. Strangely enough that curve appears on almost everyone.\page \
It's a very strange observation. We all tend, the fact of the matter is, especially in project
and engineering work, we all tend to jump in and do our best work kinda near the beginning
and then slope off. \ Now what does this say about retention and
companies? \i I
\i0 s it in fact, more the case that we should work a little bit on something and then move
to something else? I don't know but I think the data suggests that it's something to think
about.\ \
The other thing I\'92ve really seen is that everything will break a few times. This is,
this is really powerful. Seven times, I love the numerology of it. The reality of software
development is that no matter how great your team is whether it's decentralized or it\'92s
Larry Page or whatever, basically when you watch it it's going to break a series of times.
There's no avoiding that. It\'92s simply going to.\
\ The typical funny things, that happen to every
start-up, and we've got a bunch of them here in the room and I know bunch you've seen the
same failures is gonna happen.\ \
So the question is what team do you want to higher to maximize your performance in the
marketplace? The reality I believe is that you want higher the team actually not that
such specialists in this stuff but that is the most passionate about as quickly as possible
fixing it for you. \ \
And the reality is if you've got a whole bunch of people who perhaps have never, you know
done uh... CSS development before, had never worked with MySQL
\b ; \b0 the reality is they actually perform better
because they're so passionate about getting the job done and if you're basically trying
to simply get through these seven problems before your site becomes stable the question
is just how long it'll take to get there.\ \
And the reality is that people who are passionate and inexperienced or more likely to get you
there faster.\ \
[pause]\ \
We built Coffee and Power with this, can you do anything with it.\
\ Well we built this iPhone apple called Hudat.
It was Ryan, my co-founders crazy idea. It took us like twenty four hundred dollars in
about a month to actually develop and deploy. And that was built again entirely in this
system.\ \
A final thought or a couple final thoughts here;\
One is noise. This is a famous picture of basically, this picture over here on the left
is an original, in this first one in the graph here, is that picture with most of the data
taken out of it so much that you can see the Big Ben anymore. But the next pictures after
it are what happens when you actually spray-paint a little random noise onto the picture. How
odd is that? And in that second frame what you see is that optimally if you just spray
paint a little bit of random noise, suddenly you can see the image again. It\'92s this
concept called stochastic resonance. \ \
The point I\'92m making which I hope you might be able to see is that actually having a lot
of people working in kind of random ways and sometimes doing things that are a little bit
wrong on your project actually adds quality to it because it'll often discover the things
that you actually need to be doing and the things that are done wrong by those people
are easily enough fixed by the development process. So that's a fascinating thought right?
\ \
Um... the final point I\'92d make as an engineer, is people talk about scalability in the industry.
We\'92re building all these software projects and how do we make them sufficiently scalable?
\ Well the sorta singularity is on our side
there, because the final point I make about this is if you take an example like uh...
this is just a funny thought. If you look at the entire U.S. economy and all the debit
card transactions that go on it which make up the great majority of the huge chunk of
the U.S. GDP\ \
It\'92s an enormous amount of data one point five trillion dollars which actually do the
math on it nowadays you find out about twelve hundred transactions a second and the data
associated with all of us Americans buying things is less than about a megabit per second
in other words\ \
[pause]\ \
As a stretch you could probably run the entire U.S. economy on a single math book. \
\ So what I say to engineers, when I talk to
them, is if you think scalability is the reason why you\'92ve gotta hire a crack team to solve
your problem; you will be lucky as all get out if you build a problem big enough to not
be solved by really dumb solution these days.\ \
So does really matter? So I like the thought of that. \
\ So that's basically it.\
\ This is one of our great little pieces of
marketing material. \ \
So we've we built this remarkable system; using what is essentially a crowd sourced
approach. I think it's incredibly important because I suspect that the future of work,
of real, complex, design sensitive, purposeful, skilled, work, lies in the same sort of processes;
in the same sort of approach that we are so successfully seeing deployed by the gang that
is here right now. And that's the thought I wanted to leave you with. \
\ Thank you.\
\ }