Next Economy Notes 2015–11–12

James Manyika:

there are still massive skill gaps — we are missing 40M high-skilled workers 23M in china
there is a surplus of low skilled workers in advanced economies — the returns to education are very high
if you take 15 indices of gender parity and look at it for work, it is $28 trillion of additional global GDP
If every country did as well as the best country in its region, that would add $12 Trillion to global GDP
60–70% of people are no longer advancing in market income, ~30% in disposable income
wholesale automation of jobs is unlikely, but 30% of tasks in 60% of jobs will be automated
when ATMs happened, bank tellers jobs didn’t go away, but the tasks they did changed
we need to move from thinking about Jobs to Work and from Wages to Income
instead of rejecting these new modes of work, we need to make them work better

Steven Levy:

What year will computers become smarter than people?
Sebastian Thrun, you started self-driving cars at google — that seemed outlandish then

Sebastian Thrun:

I need to credit Larry Page for putting self-driving cars on streets — he asked me and I said it can’t be done
Larry Page pushed back, and I realised if a 16-year-old drunk teenager can drive, maybe a computer can too

Steven Levy:

at what point did you think about self-driving cars and the economy?

Sebastian Thrun:

for me it is personal — I have lost so many people I know to deaths in traffic accidents.
So saving lives is the most important but we can also reduce fuel consumption
when people have an accident, they learn not to do ti again. With robots when a car learns, all cars do too
once one robot makes a mistake an learns from it, no other robot will make that mistake again
it is now actually illegal for commercial pilots to fly the plane themselves in bad weather
A lot of jobs like lawyers and tax accountants can be done better by machines than humans. it will happen

Steven Levy:

we don’t trust machines as much as we trust humans — you say we shouldn’t trust people at all

Sebastian Thrun:

practically we trust machines a lot from diagnosis to banking
we all depend on machines a lot
3d printing and drone delivery will disrupt truck driving more than self-driving cars will

Steven Levy:

you started with a massive online AI course, but now you are moving udacity to lots of smaller courses

Sebastian Thrun:

we ask companies like google, facebook what it would take to get someone hired, and make courses to teach them
so we have nanodegrees on things like Android development that get people hired when they complete
we need Art Historians, but there are many more jobs for technologists

Steven Levy:

I heard that you became an Uber driver

Sebastian Thrun:

Try out becoming a Lyft or Uber driver — I recommend it to everyone to learn
In the past you needed a medallion to be a cab driver, and it was hard
Uber no longer sends me customers as a driver, so I have been invisibly fired by them
There is a company called Folded that made a game for people to play at Protein folding
the world’s best protein folder is a secretary in London

Steven Levy:

if there is one course you suggest people should take, what that would be?

Sebastian Thrun:

we have once called ‘tech entrepreneurship’ which is well worth doing
our most demanding course is android development which takes you up to Google Engineer standard

q:

my daughter wants to be a historian and is not interested in technology

Sebastian Thrun:

I get on a slippery slope when I advise parents not to send their children to college
there is a great deal of importance to studying history or art and then to bring that back to business

Steven Levy:

what year will computers be smarter than humans?

Sebastian Thrun:

In my own case 1995

Adam Cheyer:

1st there was the web, and every business needed a web site, now every business needs an assistant you talk to
viv is a cloud-based Siri or Cortana — rather than just having a few small tasks, you can ask anything
eg ‘on the way to my brothers house I need to pick up some good wine that goes with lasagna’ can be answered

Alex Lebrun:

Facebook M will handle all the mundane tasks for you, like arguing with your cable provider
M is very hard to do — it is a combination of AI and human trainers and it learns from the humans as it is used

Steven Levy:

who are the humans involved in this, are they computer scientists?

Alex Lebrun:

AI trainer is a new job, but the people are customer service people who perform the task and the AI learns from them

Steven Levy:

is tje idea you won’t need the humans at all?

Alex Lebrun:

the number of domains is very large, so we will be training the AIs for a very long time
we are trying to machine learn everything from scratch- M learns how to use websites like a human

Steven Levy:

there is a theory that assistants like these give ordinary people the chance to have a butler

Alex Lebrun:

natural language is a new UI — it as as important as keyboard and mouse were before

Steven Levy:

an you describe the personality of M?

Alex Lebrun:

we are trying to have no personality
eventually you would be able to customise your assistant as you want, but for now it is standardised

Steven Levy:

right now we make our way through the world with randomness — will these homogenise?

John Markoff:

there were 2 labs in Stanford McCarthy’s AI lab and Englebart’s Human Augmentation lab
McCarthy thought AI was 10 years out in 1962, and it is still 10 years out
designers have choices — do you design people into systems or design them out?
Tesla is running an alpha test with it’s user base now — one bad crash and it sets autonomy back by a decade

Steven Levy:

Jerry, you talk about having a robot boss who delegates to you

Jerry Kaplan:

we already are reporting to robots, adn they do a better job
we’re ebaluated by systems watching us all the time, we’re evaluated by systems for jobs and credit

Steven Levy:

you aregue we are going to give agency to these systems

Jerry Kaplan:

the issue of agency is a red herring. we need ot get the gee whiz out of artificial intelligence
if we are delegating decisions to automated systems we have already given them agency

Steven Levy:

how far are we from the movie Her and that degree of conversation?
Machines are not male or female — there is no-one home

John Markoff:

the turing test is a test of gullibility — 10M chinese are in extended conversations wiht a bot 25% said I love you
most conversations will be through proxies or with avatars
when chinese people come to america it seems very culturally quiet- chatbots are private space

Jerry Kaplan:

what I find annoying is when the machines are gratuitously friendly to me
they fool people into thinking it is a sentient being in there, and that si worrying

Steven Levy:

could a robot conduct this interview?

Alex Lebrun:

yes because I have been trained by PR to speak like a robot

Jerry Kaplan:

there is a misframing of the problem — automation doesn’t replace people, but tasks
if your job is all tasks that can be automated, then you are out of a job
pedicurists don’t just understand feet, there is an emotional connection there
so many jobs have a human connection — no one wants a robot undertaker

Kristian Hammond:

I miss the old world when everything was simple when we’d build AIs in robot bodies and they’d kill us all
Narrative Science analyses data and generates a story as structure, then we turn that into language
if the data and analysis have meaning we can transform that into language
while we would like to have abolished the whole profession of journalism, we went a differnt way
what if we could write one story for every person instead of one read by thousands
we can generate a story for the owner of a restaurant based on the credit card statement
taking a real-time data stream from IOT sensors and providing a story about it is key
if anyone has the word analyst in their job title, something like our Quill will work with them
running algorithms against data gets new meaning, but if the meaning is there is can be transformed into language
for me, reasoning is not magic, but these systems need to explain themselves to us
I said a robot journalist will win a pulitzer in 5 years 3 years ago
i don’t do a lot fo work in journalism any more

Steven Levy:

not paying enough?

Kristian Hammond:

there will be pulitzer quality stories, but we will not write them, someone will use quill to do them

Tim O’Reilly:

Kickstarter is an example of the new approach to funding from the crowd
Clay Shirky said we moved from filter then pubish, to publish then filter, and Kickstarter does that for funding

Yancey Strickler:

I want to talk about what is possible when you are not forced to be in the business of just making money
There are 25,000 projects done by people who have already created projects 75% success rate for subsequent ones
hardware makers used to need massive scale — before 2012 we were either handmaking or 100,000+ runs
that change to where someone can think about what they are doing fro 6 months and make 150 to 50,000 units
Boardgames are one of the healthiest communities on kickstarter — $250M in aggregate, lots of small games
Kickstarter is self-selecting groups of communities producing new kinds of creativity
can hardware be an artful medium?
on kickstarter ideas are funded because people think they are cool, not because they will make a profit
they legalised crowd funded small investing recently, but Kickstarter is not going near that
at some point in the last 15 years it became OK to sell out. Don’t Sell Out.
It’s very simple — we spend less money than we make at Kickstarter
If you keep your independence you can call the shots. This is a challenge for creative people
creative people often feel guilt about making money, and underpay themselves
You have to be idealistic and believe that it is possible to change that you aren’t just screaming into the wind
Patagonia, — a benefit company — says they will share proprietary information with competitors to help the environment
basecamp, @pinboard these are self-sufficient companies that care about what they are doing
Fugazi i see as punk rock heroes but they are also small business heroes who’ve kept it going for years
Harvard Business Review sat to be paranoid, disrupt yourself, go to war — all very toxic ideas
This language is anxiety inducing — it makes you want to behave badly
it’s harder because your tools of measurement are going to be handwavy, but easier as you can be principled
Kickstarter became a Public Benefit Corporation — optimising for benefiting society, not shareholder value
we are going to publicly report on what we have done to benefit society
it is not a purity test of ideology, it is polyculture of many different ways of thinking
we champion and celebrate the creation of art and culture. Fuck the monoculture

adafruit industries:

Adafruit is an open source hardware company based in manhattan
I design here in the factory and the manufacturing is right here — I have pick and place line in my Soho factory
I can design and manufacture 5 seconds apart

Tim O’Reilly:

I believe you are the last cellphone manufacturer in america

adafruit industries:

we have an open source cellphone that you can make yourself
components come on reels, they are picked up 12 at a time and put down on the board at superhuman speed
we have 1 50,000sq ft manuafacturing facility in Soho, about 100 people, grossed $33M last year
we haven’t taken any VC, we are wholly self funded and build out on our own way
I grew up in culture that worshipped open source, and I have brought that over to the hardware side
we have a hangout every week for creators, and we have ‘ask an engineer’ live show
a 6 year old who was watching my show asked her father ‘are there male engineers too?’

Tim O’Reilly:

open source hardware is a big part of what you do

adafruit industries:

I have 500–600 github repositories, one for each project, and objects on thingiverse
we have a community that remixes and rebuilds what we make.
we’re not going to do patents for the sake of it 17 years is too long, it changes in 6 months

Tim O’Reilly:

you said your pick and place machine documentation was terrible so you made video tutorials to teach it

adafruit industries:

last year we launched our learing site which has over 900 tutorials from stripping wire to making a laptop
good information is advertising — instead of buying ads we put up this inofrmation that leads people back to us
we’re an education company with a gift shop at the end

Tim O’Reilly:

there are multiple ways to monetise — if you spread education and knowledge you find a way to make money
I remember you having a strong opinion about different kinds of resistor or special vacuum tubes

adafruit industries:

electronics is a form of art, it is the artform I have chosen to express myself
historically electronics people have pushed fashion aside, but now we can bring them together
I went to a fashion show and sow clothes with LEDs in with the code written by girls from the Bronx

Tim O’Reilly:

there’s a choice not just to be consumers but to be creators

adafruit industries:

now we see 3 year olds with iPads, but people don’t know what is in them_what if you make your own with raspberry pi
if you have someone who you want to help, find a small project from our site and do it with her

Mark Hatch:

I love revolutions — I am a former green beret and I love blowing things up
I get asked “has anything serious come out of the maker movement?”
when you put the tools of the industrial revolution into the hands of the middle class they do amazing things
this is the worlds fastest electric motorcycle — it was built in our Menlo Park techshop
Andy Filo is working on a jetpack — he is now consulting to NASA for their jetpacks
Michael Pinneo is building a desktop diamond manufacturing device 95% hydrogen, 5% methane create plasma, diamonds
Tina Albin-Lax came in and learned how to use a laser cutter, and cut out cupcake toppers for her nephews party
at the end of the party she had $400 worth of orders and then it became a business
David Lang wrote ‘Zero to Maker’ — he took 20 classes and 9 month later started an underwater robot company
Brit Morin came to Techshop, learned a bunch of tools and now has a huge lifestyle craft business
Mark Roth was homeless, he came into Tech shop and built Tchotchkes and sold them, and built a business
James McKelvey made the prototypes for Square in our Menlo Park TEchShop
the fundamental question is when are you going to join the revolution?

Saul Griffith:

July 1945 Vannevar Bush wrote ‘As we may think’ describing a vision for what we could be
in the same month he submitted to the president “Science, the endless frontier” which became the NSF
research is changing a lot — Google spends $9B on R&D NSF is $6B, DARPA $3B this is a big shift
we have a solar powered plane that generates 35 KWh and flies forever
We have company called Sunfolding that tracks the sun to get 30% more out of solar panels
we have a precision pointing device made out of compressed soda bottles
the department of defence gave us money to build a walking inflatable elephant that walks 1mile in 8 hours
traditional industrial robots have to be kept in a cage, we’re from california for cage-free robots
our pneumatic robots are the first ones that can lift their own weight
our elephant inspired Big Hero 6 — the future of soft robots
we built an exoskeleton that makes people run faster with less energy
I want this to run a marathon without training, but it’s real purpose is stroke rehabilitation
For a million dollars you can now have very high precision measurement and manufacturing
so we will have a lot more independent labs, and more diversity of labs and people building them
we’re not investing in solving the problem we know about — Energy

Tim O’Reilly:

the company founded by Edison, GE is also thinking about thse issues -here’s Jeff Immelt

Jeff Immelt:

I’ve been at GE for more than 30 years. For most of that era the role of CIO was not significant
if you look at the innovation in the consumer space, most of that hasn’t made it into the industrial sector yet

Tim O’Reilly:

You recently appointed GE’s first Digital Office, but you got rid of the finance unit that was more digital

Jeff Immelt:

We’re $150B of making things — we are the worlds biggest industrial company
a locomotive today is a rolling datacentre — it has 600 sensors of fuel efficency and motion
for decades CIOs were doing the help desk, and seen as a cost centre. We want Digital Industrial

Tim O’Reilly:

will this industrial internet be open or closed? is it open source, is it like the web?

Jeff Immelt:

I grew up in a closed world, but it now needs to be an open system. That’s not where I started from
we do high tech big systems — Jet engines, turbines. Labour cost is 10–20%, materials science matters more
manufacturing used to be an out-sourced labour market arbitrage business
now it is more high skill and you are designing the manufacturing process as you design the jet engine
Silicon Valley doesn’t get that people and culture are even more important in big companies than small ones
when you see an airplane engine that is 20 years old, you want the people who made it to still be working there
we acquired a company called WorldTech that does cybersecurity, and we work to harden the electricity grid
the number of manufacturing workers in the US is growing even as we improve productivity
we are investing as much today in new material science that our supply chain can’t do, there is a time lag

Esther Kaplan:

average earnings is $35k a year, but retail is $20k/year
within days or weeks of companies adopting electronic scheduling system we see changes in people’s work

Darrion Sjoquist :

My mother working at starbucks meant i had to get up at 5am to get my sister to school for 7am
my mother would start at 3am and my schedule would depend on hers and I then had to take my sister to school
I could not sign up for any clubs at school because I could not rely on my schedule being stable over time
When my mother moved to a secretarial job with regular hours I was finally able to take up theatre
Now I work at Starbucks too. We are perpetually operating at the minimum number of people possible there
if you call in sick, you know that your co-workers will go through hell because there is no slack
Starbucks said they would give people more advance notice 9 months ago we started seeing change 2 months ago
while there is an official end to clopening (closing the store and reopening the next day)
we still see people leaving at 9.30pm and needing to be back the next day at 6.30am
for a shift supervisor, the concept of calling in sick is impossible — you can’t get someone to cover it
I’ve seen barista’s come in with flu or pinkeye because the system doesn’t allow for people to get sick

carrie gleason:

I have seen a shift to part-time employment since 2005 — people with fulltime jobs were asked fro more shifts
then shifts started to change from week to week, and employees had no input, the managers set them
41% of all hourly workers get their schedule with under a weeks notice
when you don’t know your schedule from week to week, you can’t plan childcare or your life
workers are using facebook and setting up shift swapping groups to exchange shift hours
in san francisco workers are paying each other money to cover their shifts

Esther Kaplan:

people get scheduled for shifts with a start time but an undetermined end time

carrie gleason:

victoria’s secret workers will get 4 hour shifts, then scheduled for 2-hour on-call shifts afterwards

Esther Kaplan:

in effect your whole life is on hold to get these 10–12 hour shifts each week

carrie gleason:

people who try to work retail and go to school at the same time end up being forced to choose between

Esther Kaplan:

the American myth of social mobility is being interrupted by these scheduling models

Charles DeWitt :

You mentioned that wal-mart redefined retail in the mid 90s — it was a mindset change
labour was seen as a bundle of costs and liabilities through ERP, and not the value of continuity
in 2009 the national retail federation show people asked how to get more out of labour
if you’re good to employees, they will be good to you — if it is too lean you lose out
a lot of longitudinal studies show that engaged employees are better employees
firms are now asking us how to improve this — we need to model fairness and make it visible so they can fix it
we are working to move from transactional systems to reflect these variables.

carrie gleason:

When I first started seeing these trends happen, the schedule was not created by the manager, but software was
then the manager role became cost containment — they could only show growth by shrinking labour
the managers would skim off hours — get sent home early, get hauled in last minute
the managers would blame the software for the schedules, but they were froced to squeeze the workers hours

Charles DeWitt :

10 years ago ti was all about cost containment and coverage, but the algortihm didn’t consider hour stability
scheduling is about getting people there are the right time — it is a policy and configuration issue
there are mathematically right answers to allowing for sickness, but it needs to be built in

Esther Kaplan:

the interplay between turnover and staffing is not visible on the same chart or to the same people
is the technology there to show the full impact of this lean scheduling?

Charles DeWitt :

there is an issue of it being a siloed organisation and information not flowing
in our system, all of this data is here, but it is in a transactional database, not a reporting database
Marsh Fisher wrote ‘the new science of retailing’ showed the correlation between labour in store and sales
the CEO says ‘you need to cut 5% labour’ “lets close 5% of stores” ‘we’d lose revenue’ “exactly”
the retailers I am talking to do want to change, so we can feed back this data to them to change

carrie gleason:

these problems have been around for a decade; legislation swept across the country
now 250,000 workers no longer have on-call scheduling — as retailers reform practices

Darrion Sjoquist :

the reason that I am here is to make it apparent that we are people and we need consideration

carrie gleason:

@fairworkweek are launching the High Road Workweek partnership today
these practices have impact on community, economy and individuals
predicatability, adequate hours, employee input and flexibility, equal opportunity and mobility
Technology has a critical role to play in this, but the company needs to set the values and measure it
managers need to be set up for success with metrics aligned to scheduling equity principles
go to fairworkweek.org to see this pledge

Lauren Smiley:

Respond on medium and tag with #wtfeconomy to discuss this
Leah Busque founded Taskrabbit long before Uber or Airbnb existed
when you started taskrabbit in 2008 it was a very different environment how did it come to you?

Leah Busque:

I was a software engineer at IBM for 8 years. I had the idea when we realised we were out of dogfood in the snow
wouldn’t it be nice to ask if anyone nearby could pick up the dogfood and drop it off
in 2008 the iphone had just come out, and facebook was only just getting out of colleges
location data was not being used then. I came up with this idea as ‘service networking’ not ‘social networking’
when I launched taskrabbit in Boston Iived in Charlestown and I thought I’d get lots of college students
I recruited 600 moms in chralesetown for the beta network; I thought i’d get students as taskers
I din’t get students, I got moms who were out and about, and retirees, and young professionals at weekends
out of the 30,000 taskers the top 10% is doing it full time — 3 tasks a day, ~$7k a month the rest are casual
a lot fo the most successful taskers are handymen or housecleaners who had a background in that industry
we have been in the market 5 years — when we started there was just craigslist
after 5 years we saw the landscape shift — a lot of companies were specialising in differnt things
we stepped back and said ‘what if we were launching now?’ and rebuilt in 2014 to focus on mobile, scheduling
in the previous product there was an auction bidding model that would take 24 hours to resolve like ebay
now we can schedule it more practically in advance — the taskers set their rates and availability & show up at once
the great thing about this new platform is that taskers are making more money as they set their rates not bidding
we tested this in London first before rolling out in US and taskers made 400% more monthly

Lauren Smiley:

you did lose some people after the change?

Leah Busque:

some people loved the auction/bidding model but most who made money make more wiht the new system
on average taskers make $35/hr and overall about $900/month — so part time
that we allow taskers to set their own hourly rates, schedules and accept tasks on demand
we have to make sure that both sides of the market are served by the system
home services has been a very fragmented space in the past — we have put together a way to do it sytematically
taskrabbit is a technology platform that enables this new generation of workers
taskrabbit.com is one place that could generate demand, but amazon can generate demand for installing TVs

Andreas Weigend:

how is this different from what we were hearing about Starbucks et al in the last panel?

Leah Busque:

Many new companies sacrifice the supply side, the worker side in the race fro growth
I think this model could be applied to other companies and industries

q:

you do criminal records for the taskers but not for the people soliciting the tasks, how come?

Leah Busque:

you have to think about the expectations and needs of the supplier and the consumer
we ask for credit card upfront before a job can be posted to verify address and credit for the consumer
we could do more there, but what we have in place has worked well so far

Jon Kessler:

I was a cab driver, and my income was declining rapidly, there was a lot fo downtime. so I signed up for Lyft
before Uber and Lyft being a cab driver was good, passengers always said ti was hard to get a cab
there are a lot of drivers who do both, but the cab companies are very anti Lyft/Uber
I did not own a medallion — I did not purchase one

Kelly Dessaint:

I started doing Lyft in Feb 2014, the pries were good, but the rates charged kept going down — 40%
I found it wasn’t good for the drivers — you’re using your own personal car as a cab in SF
I did Lyft for 10 months and my back seat looks like I transported farm animals

Lauren Smiley:

you went to cab school — how many others were in your class?

Kelly Dessaint:

one, but he went back to Lyft
taxi drivers are a community — we talk to each other. We all communicate through GroupMe
we talk on GroupMe about traffic and closed streets and where pick-ups are good
Lyft and Uber drivers are horrible drivers — very few live in San Francisco and they drive like tourists
I’m on Fell and a car is coming at me and all I see is a pink glow-stache and I hope he stops in time
I work public works Friday Saturday night on the cab rank, and I see 4–5 uber/lyft

Eric Barajas:

I started part time driving Uber for a week, and I was able to make my own schedule. compared to my last job
I saw an ad on craigslist saying Uber was $35/hour and I didn’t make that then
I’ve been at it 9 months and I feel trapped — I work 11 hours a day+ 2 hours commute
gross I make about $1050 a week — the IRS deduction is 57¢ a mile so about $175/day gas is $40–50/day
my car lease from Uber is $170/week. Insurance was charged per mile and it came to $540/mo

Jon Kessler:

my expense for car and insurance+ phone are less $488/mo — I drive between 7 and 30 hours/week
the more hours I work the more I make — on average about $28/hour net before tax

Kelly Dessaint:

I have no expenses besides my gate fees — thursday $111 friday $121 saturday $111 — that gets me car, everything

Eric Barajas:

I like the flexibility and the app most, but I don’t like how they treat the drivers

Lauren Smiley:

I met Eric at an Uber protest outside the company

Eric Barajas:

what we want is tips and higher rates
airport is about $32 each I’ve never seen a surge from the airport, $50 from union sq to airport

Kelly Dessaint:

For a taxi, the airport is $55 so we don’t get so many of those
at national we put out all the cars, deSoto puts fewer cabs out
we have specific guys who come in adn do the short trips
my friend Late Night Larry says you just say “you’re out” and throw people out of the cab

Eric Barajas:

I fear deactivation — you can get instantly fired

Kelly Dessaint:

with cabs you get called down to the MTA to answer complaints, but no-one ever gets called in

Jon Kessler:

a cab driver has more power than the uber/lyft driver as you are the cab companies customer
uber the passenger has all the power; lyft is more balanced; cabs the driver has it

Kelly Dessaint:

if you are in a situation that you do not like, get out of the cab, keep it between you and them
you have a right of refusal over passengers -

Jon Kessler:

you can refuse if they leave the city, if you don’t have time, they’re dirty or belligerent
you are supposed to pick up drunk people so they don’t drive

Kelly Dessaint:

there are so many reasons you can not pick up a hail, you still have that control
if you don’t pick them up, they don’t have much sanction

Tim O’Reilly:

we think about these platforms from the user point of view, but what does it say about corporate structure?
Coase said that the management overhead of the large firm was less than the transaction costs between small firms
we are starting to see small core companies with a large network
Esko Kilpi reminds me of a conference I did called ‘building the internet operating system’ before cloud existed
I found someone then who was unknown — Clay Shirky- Esko reminds me of that experience

Esko Kilpi:

Lets pretend that we are in a conference about lifting things call it Next Lifting
imagine we are discussing when the cranes will be able to lift as much as a human
people are not clever, people have never been clever — we need all the help that we can get
technology is a vital ingredient
How can it be that we are more productive after retirement?
the work systems we have are not only broken but are based on artificial scarcity
when we talk about algorithms they are very different from how humans behave
algorithms are discrete; humans are not discrete so modelling a human is impossible for the forseeable future
we should understand the algorithmic economy as an absolute necessity, and it requires matehmatics
in the old system your boss told you what the problem you are supposed to solve it
in this new environment the problem has not been defined beforehand; work starts with problem definition
we need to allow people to find ways to define problems, and ways to solve them by collborating with other people
we need everything to be decentralised; without a network it won’t work to solve problems
the thinsg that are platforms now will become a commons in future
work and learning have to become the same thing

Steven Levy:

we now have a panel of VCs

simon rothman:

I invested in Sprig — a startup that will bring food to you within 10 minutes, not from a restaurant
it’s obvious that I can press a button on my phone and have food brought to me in 10 minues

Gary Swart:

the single biggest lever companies have is talent and it is hard to get that oDesk, now UpWork is exactly that
it is B2B SaaS company

Steven Levy:

can you explain what that means?

Gary Swart:

it’s very hard to hire good sales people and ti takes 9 months to know when they got it right

James Cham:

one of the 1st investment we made was in a company called fieldwire that is construction software
when the line cook has an android phone that can take pictures of anything, everyone is a knowledge worker
there is a set of people who are introverts who are focused on tools that make tools — software developers
if you apply tools like make files to otehr fields lots can happen
if you’re a software developer fieldwire looks like Jira, but it makes a difference in construction

Gary Swart:

we get excited about innovative technologies, but also about new business models
in a pure marketplace it is a beautiful thing, but many so-called marketplaces are managed services
then the goals are misaligned — you want the lowest quality at the highest price, the customer wants the opposite

simon rothman:

many people look at uber as if it is a service, but actually it is a marketplace
regulation is an historic artefact — it models the past. I fyour idea is so big it will require new regulations

James Cham:

silicon valley is test lab fro the rest of the world

simon rothman:

in the early days of a startup you have to beg the media to write about you; when you grow you can’t stop them

Steven Levy:

are these marketplaces essentially winner take all? the way you merged companies

Gary Swart:

if not a winner take all it is a winner take most — look at ebay buying Stubhub — you see some verticals

Nick Grossman:

I learned a lot on the internet, and thank you all for that
I work at UNion Square Ventures, and we invest in networks — many of these help people become professionals
you can become a recording artist at Soundcloud wihtout a record label ,a product designer on kickstarter
in the old days there was the firm and it defined what a job was
now there are lost of smaller possible firms
a lot of thinsg are attached to firms — healthcare, tax collections, unions
these arenow becoming unbundled fromt he firm and attached to the individual worker
how can we rebundle these services around the worker in new ways
job discovery and scheduling; benefits and insurance; identity and reputation; community and organising
education and training; facilities and equipment; finance and admin
there are startups in all these sectors
who are these companies working for — the worker or the firm — we want the worker as customer
wheneveryon is their own tiny company, the benefits need to attach to the individual, not the companies
many startups end up as identity and reputation platforms — can on-demand workers move their identity around
in the networked economy, can workers have leverage?
the 2010s have been the decade of networked activism — we haven’t seen this yet for the labour market

Steven Levy:

Stewart Butterfield is the founder of Slack, a way of improving conversations within firms

Stewart Butterfield:

Slack is a messaging app for teams
most people have used ICQ, or AIM or IRC — messaging is a classic tech tool, but it is new to companies
at another O’Reilly conference 11 years ago when I introduced Flickr Ben Ceverny talked about tarnstions
computing used to be about applications; then it was about documents, now it is about relationships

Steven Levy:

Slack looks like it coudl eb a consumer product — it has whimsy in it

Stewart Butterfield:

we have a weird origin story -the company was founded to build a multiplayer game
we had designed the proto-slack to deal with our internal communications building that game
it was built incrementally to solve each problme in the smallest way possibel as we encoountered it

Steven Levy:

facebook was not at all the first example of socisl networks, they just got the dials right

Stewart Butterfield:

1.7 million people use Slack every day around 500,000 paid subscribers $8/mo $80/year
using Slack for free is like a free trial that doesn’t expire
when ti is free you only have 10,000 message history and 5 integrations — by the time you hit those you want to pay

Steven Levy:

Slack is a very sneaky application — you use it naturally and gradually it takes over your work life

Stewart Butterfield:

what slack does is make lots of the internal processes transparent
when you join a team that is using Slack you can see the historical archive at once

Steven Levy:

Slack is a subversive way of opening up an organisation, whetehr they want to or not

Stewart Butterfield:

there are things people can do to hoard information, and Slack makes that a little bit harder to do
this is a transformation that has been going on slowly for the last 20 years or so where you get more access to info
It was always a fact that howard street was 1-way in that direction, but now I can see that by pinch zooming
we don’t even notice how much more powerful we are as signal processors as human beings over the last few years

Steven Levy:

slack sucks in information from all the other apps too is it the one app?

Stewart Butterfield:

slack becomes the one app you have open all day alongside whatever other app you have open like excel, photoshop
we get about 15–20,000 tweets a month and they are all fed into the tweets channel
anyone in the company can look at them, before only support people woudl see that
we also make it a rule that everyone in the company has to do support too

Steven Levy:

are there plans for a consumer slack?

Stewart Butterfield:

we have no plans to do a consumer slack. A slack-like version fo Ning woudl make sense
we do see a lot of family-based use of slack, at least in san francisco

Steven Levy:

we use slack with 1099 workers at medium — they have access to some channels

Stewart Butterfield:

we do not currently do a good job there with 1099 workers, but we need to do better
at 8 people everyone knows everyone else really well; the amount of communication doesn’t scale up
we were 80% public when we were 8 people and at 100 people and it’s now 70% private
the challenge is that if all of it were public peopl would ahve to read too much
there is value for a question to be asked and answered in public, but people can be hesitant about it
we can give advice to people administering slacks,

Tim O’Reilly:

the organisational unit for military for 1000s of years has been 8 people and that is a good size for slack groups
why are so few people self-employed? 11.4% in 1990 10% today — why is the gig economy not showing up?

Brad Smith:

when we started it looks liek about 6% were self employed — we now see 36% or so
the Pew study measures people who own or are partners in a business — we measure the contingent workforce 50M
where the 2 numbers get out of sync- 79% of the self-employed work part-time as well, so that gets counted
if you expose the government to the data you can get them to change their thnking

Tim O’Reilly:

we work within these rules that were made up a lmng time ago and we don’t think that we can rewrite the rules

Brad Smith:

our entire system has been set up for people who have a regular paycheck
in fact we find most people we see have 3 sources of income, and this doesn’t match regulaion
today if you’re self employed and have 3 sources of income it is hard to know how much money you have
QuickBooks self-employed lets you say which belongs to personal or business expenses
when you’re working for Uber, Lyft etc you are actually a small business, and we can file your business taxes
The government requires paper copies of receipts as a business now
our model at intuit is to find the biggest unsolved pain and solve it — tracking business vs personal was that

Tim O’Reilly:

Anne Marie Slaughter in “unfinished business” talks about how much is left to do to equalize men and women

Lauren Smiley:

when your story in the Atlantic came out it slayed me and all the women I know

Anne-Marie Slaughter:

before I went to the state department I believed that if you want it enough you can make it work
I thought that you can make work and family fit together, and realised that you can’t
people then thought that I wasn’t a player in DC because I put my family ahead of my work

Tim O’Reilly:

we really undervalue care in our society — we have a masochistic work culture
we have one class of people who are entitled to work harder, and another class that we put out of work

Anne-Marie Slaughter:

this is a US problem. Other countries have an infrastructure of care — Sweden invests in the next generation
in this society the fastest way to be devalued is when asked ‘what do you do?’ to say that you are a carer
if you don’t take care of your families or your workers you are paying an economic price too
we are now below Japan comapred to the fraction of women in the workforce
now daycare costs more than rent in almost all states so people stay home instead
we pay people who care for our children the same as we pay people to park our cars that is incredible
For parents and families the on-demand economy is a godsend. Care is unpredictable
one ear infection can destroy a weeks worth of meetings
when you can schedule your own work that is a solution to the care problem
my son drew me as a laptop, not a person in a laptop, and my husband cooking.
even if companies offer parental leave but the senior men don’t take it, it doesn’t make a difference
in my foreign policy life, we see large groups of unemployed young people as the ingredients of revolution
loss of family life leads to political turbulence and dysfunction, so we need to support it
tomorrow the ‘good work code’ is rolling out
these are work issues, family issues, middle class issues, not just women’s issues

Lauren Smiley:

was it hard to admit that you were not the lead parent?

Anne-Marie Slaughter:

My husband said “you’re not here” — parenting is a full time job and one of you is the lead. Value that
by day I’m a feminist, by day I’m the worried mother of 2 boys — “when did girls get smarter than boys?”
looking globally women are the vastly untapped resource
in America women go into the workforce and do well until they are sideswiped by the caring issue
we have a whole debate about short-termism and how it hurts the long term bottom line
I think we are going to have to make legal changes to fix the tension between employment and caring
the ability to live at home and be supported at home connected to a network of caregivers is going to be key

Originally published at www.kevinmarks.com on November 12, 2015.