Video Of The Week: TechCrunch Disrupt Interview
Ron Conway and I did an interview with Kim-Mai Cutler on stage at Disrupt a couple weeks ago. Here it is:
Ron Conway and I did an interview with Kim-Mai Cutler on stage at Disrupt a couple weeks ago. Here it is:
Comments (Archived):
I agree with you about sharing data. It’s an emerging and under-appreciated trend.1/ Can be used in the aggregate to counter ill-fated public policy, and make a better case for progress.2/ Can be used individually to gain points, get discounts, earn services, etc. when it’s built-in as part of the value delivery of a product, eg. Adjusto is an insurance App in Canada that earns you insurance discounts if you share your driving habits http://www.ajusto.com/Healthcare is yearning for data sharing at the peer to peer levels.Take my data, please!
Yup !If one visualizes social systems at all levels as living systems then collective data ownership/awareness of each living system’s appropriate level-matched data is fundamental to its survival strategies.I need access to my life data !as We need access to our life data !https://www.youtube.com/wat…
Im as much a data person as anybody, but while the common group think is that data is clean enough to support (fail to reject) a hypothethis, most data is noisy, subjective and misused. Moreover, watch what you ask for! once that data and its results gets in the hands of a politician, look out!
well..needs to be qualified:1/ structured data2/ data with value in it3/ data for a purpose4/ not private dataetc…
Not sure what you mean by structured data or the rest of the list, there is supervised and unsupervised data (difference is the the existence of a dependent variable).note: data can have no “structure or even (perceived) value” but still support a conclusion.
I agree data can have no structure or even (perceived) value but still support a conclusion.I also get what @wmoug:disqus is saying with his four types of qualified data.The way I break it down is:Examples of structured data———————————Get user to input information into a box that’s pre-labeled (aka meta-tagged), such as:* name* date of birth* occupation* genderExamples of unstructured data———————————–Get user to upload information that has no prior labels, such as:* freeform sentences in a textbox* image file* sound file* video file* digital bookWrt health records, the structured data would be the patient’s name and health insurance number.The unstructured data would be the doctor’s scrawling hand-written notes of diagnosis.Supervised learning———————–Generating a link between two data sets where DataSet 1 is the observed data and DataSet 2 is the predicted outcome.DataSet 1 is often clearly labeled, e.g. X-axis: incidence of cancer = 1, no incidence of cancer = 0; Y-axis : age of patient.The hypothesis then is something like: “Given the last N observations provided in DataSet 1, calculate the likelihood of someone aged 35-40 getting cancer.”Unsupervised learning————————–Data is less clearly labeled/formatted.The intelligent algorithm starts by first classifying it by type (images = jpeg, png, gif, file size ; video = time duration, file size; sound = mp3, file size; words = word count, frequency of words).Then it may apply any number of probabilistic and statistical methods to establish correlations and common frequent occurrences of:* words and phrase structures in the text document — how text mining works* recognizable items in the image file (e.g. cats, trees, cars)* phenomes and other voice patterns in the sound fileYour comment “most data is noisy, subjective and misused” is where it matters.The assumption for hundreds of years has been that probability and statistics can deal appropriately with subjectivity.Regardless of whether the data is structured or unstructured and whether supervised or unsupervised learning is happening, this underlying assumption exists.The question is, “What if probability and statistics are NOT the mathematical tools by which to qualify the data?”How do we define the value (individual and comparative) of an atom of data?Is the data about a person’s age more valuable than the data about a person’s eating habit in determining their health?The issue with sharing health data on a p2p basis is, of course, tied to insurance premiums and employment agreements, and to what extent third parties who aren’t our doctors have access to that data.
Your list suggests you mean information, not data.
Well, I mixed objectives with data type, but we’re talking real data whether aggregated or anonymized.
With no data to back this up I feel that we have a transportation issue and not a housing shortage. From my knowledge is seems as though people want to live in an area because of the amount of time it takes to get to the area.
Yes. High speed (and safe) trains FTW.
Lots of people in Connecticut commute into NYC, some from ridiculous distances. When I worked in New Haven everyone talked about the “gold standard” being their desire for a 1 hour direct train to and from NYC. I can only imagine what kind of boost upgrading rail transit to the city would provide our economy here.Up in the Hartford area we just brought online a bus rapid transit system and so far it’s been more successful than all the models/predictions.I also think WiFi/LTE is huge in this regard. If you’re able to work during the hour you commute it’s not a lost hour the way an hour driving an automobile is.
Tethering makes my commute bearable. It’s a big plus. Really makes a world of difference.
everyone talked about the “gold standard” being their desire for a 1 hour direct train to and from NYCWhat is the true portal to portal time though of a “1 hour train”? For example you still have to get to the train station, wait for the train (you aren’t going to arrive exactly the minute it leaves..) and get from the train to where your office is. And then do the same thing in reverse. Most people don’t live and works in train stations.I would think that a true 1 hour train means a 1 hour 20 minute to 1 hour 40 minute actual commute.
“If you’re able to work during the hour you commute it’s not a lost hour the way an hour driving an automobile is.”Yes! I used to get my best work done on the train to Providence in the quiet car. I looked forward to it.Sometimes I wish we could trade all this work on self-driving cars for investment in more trains.
> Yes! I used to get my best work done on the train to Providence in the quiet car. I looked forward to it.I had a much different experience. The amount of time/day dedicated to traveling in addition to a “work day” was too much for my life.> Sometimes I wish we could trade all this work on self-driving cars for investment in more trains.I don’t really understand this. Why do you want that?
> Up in the Hartford area we just brought online a bus rapid transit system and so far it’s been more successful than all the models/predictions.Is this information available online anywhere?
http://www.courant.com/news…Lots of other articles if you search local media on Google as well.
Thanks – I’ll look around for more recent reports, too.
If there is demand for it, and people will pay for it, they should build it.
What do you think of ridesharing and car pooling to solve that?
I don’t think so. From my knowledge and experience most people don’t want to car pool or ride share. Also, those options are only good when you can be flexible and understanding. Sometimes you can’t.
This is where self-driving car will forever change the economy over next three decades. Have written on this in detail before, but short version. Self-driving car (and all the resultant decreases in travel time / cost / efficeicny) will essentially double the land area within same distance of work places (say 30 – 60 minutes which are big cut off points), thus result in a 4 fold increase in housing supply (area is pi * radius – which has squared – doubled). So basically 50 miles from san fran will be a 50 minute trip vs nearly two hours, which will enable 500k-1M houses to be built for 100k – 300k… completely changing the landscape. Probably 2 decades off, maybe only 1. Single worst investment is residential real estate priced above $150 square feet (most tier one cities are between $400 – $1,000 square feet), think wisely 🙂
> Self-driving car (and all the resultant decreases in travel time / cost / efficeicny)Do the changes in time, cost, and efficiency assume that all vehicles are self-driving?Where do these numbers comes from?
Would it hurt credibility to say from my head? 🙂 I say that partially and jest and in reality. I have studied this issue pretty extensively (100’s of hours, whether that lends to credibility is in the eye of the beholder), and the key number around price per square foot is that you can build new single family and multi-condo (non high rise) for between $75-125 pretty reliably. No doubt many if not most new homes are sold at higher amounts, but this is not due to construction costs but instead geographrical constraints (only so much “prime land”) and zoning / regulatory barriers (such as minimum acreage requirements) that make land value scarcer. So from that stand point, once self-driving cars make being 50 miles from san fran, new York, Chicago etc. not a big deal eventually costs should trend to marginal or true building costs, which themselves can / should go down over time. Not sure if this is helpful or provides anymore credibility, but this is the basic argument, one which I think portends the future. In regards to if all need to be self-driving cars I would definitely say there is no need for 100%, but I would say that probably north of 50% of cars would need to be self-driving for the effects of “expanded real estate proximity” to be meaningful, but I think once self-driving cars gain broad legal / regulatory acceptance that 50% threshold will be crossed quickly. Getting to 100% may never happen or may take multiple decades as there can be some holdouts etc, but a meaningful amount of self-driving cars will happen quickly after acceptance as the economic case is so compelling (particularly with uber / lyft networks cost per mile can be reduced by orders of magnitude)
I think one of the big issues in the NYC area is capacity. I don’t see how automated vehicles will fix that problem. I think they may help. I also think that any improvement will cause more people to want to commute and it will cause an increase in travel time.I think there either needs to be a new type of transportation infrastructure or major investment in the current infrastructure to increase capacity.
One quick example is self-driving buses. Basically with drivers removed buses can be about 70% cheaper per trip, and wait times get driven towards insignificant. So this will make people be able to travel much quicker to where they want to go. So in some ways it will have the effect of making current infrastructure be able to move 2x+ more people. One recent study (I think from MIT?) used the example with mass transit / buses in boston being able to move 205k people per hour on a highway vs. current 40k. So basically a five fold increase in capacity for no cost, and in all practicality demand / miles traveled will not increased five fold so you will have more movement WITH less congestion. Again, certainly probably a decade to two away, but almost inevitable eventually in my estimation. Hope you don’t own overprice real estate in the city! If so, pay it off or sell!
Transportation and housing are interlinked.
The Tech ecosystem does have a responsibility to get involved in society at large, of course, but the biggest impact the Tech industry can have is by relentlessly continue to innovate, IMOHO.
yup
Instead of saying “low to middle income,” we should just say, “the majority of people.”San Francisco was a fascinating place to live. I think it may be the most civically engaged city in the country. It’s definitely a part of the culture. The population’s willingness to blaze new trails (for better or worse) is a service to the rest of the U.S.If the tech community wants to improve civic involvement, they should make election day a paid holiday.
If I had to bet money I’d bet that a majority of tech workers are voters (at least those that are qualified to do so).
Middle income is better than ” middle class”.
Agreed. My point is that using either one of those terms conveys that it’s a minority or subsection or special group. It’s actually practically everybody. I think that saying “low to middle income” in this context “others” the problem when in fact it’s a problem for most people.I think that Conway, by using the term, was referring to the problem in SF whereby most of the new housing being built is for *high income* people. But that’s the minority group.
The neural-net that is human commercial activity has long been glued together via somewhat universal pricing signals.That glue is now starting to seriously fail under the weight of organic network-effect interdependency conditions.Open data is no longer an option but a survival necessity.
Ron stated that 10 years ago (2005) the tech community was a tiny percentage of “any cities population”. Then he says that “today the tech community is a much larger percentage of the population of the community, so tech has to get involved in the community”.I don’t believe this statement is any where near correct for “any cities population”. This idea that tech is a large part of a typical American city (large or small with only a few obvious exceptions) is simply not correct. Sure, it is “larger” (whatever that means) but it is for sure still a very “tiny” portion of the average American community. Once again, America is not SF Bay area and/or Silicon Valley.
It’s interesting how when you were asked if Tech is responsible for the impact that they are making (to the negative) “yeah it’s going to be really hard” (disruption and layoffs) you answered in a way similar to what the mafia does (in popular culture). More or less a Robin Hood response “rob from the rich, give to the poor”. Give out turkeys at Christmas and do a great deal of good helping people to nullify all of the critics and create alliances that prevent various important parties “keep your enemies closer” from feeling that they are not getting enough of “the cheer” and therefore there is no upside “so let’s crack down and pass laws”. [1]What’s interesting is that that is one of the upsides of concentration of power. If the disruption was happening w/o a concentration of power (split between millions of people benefiting from the disruption) there would certainly be less of a chance of any one person actually feeling guilty enough (needing to wash their hands) to actually donate or do anything to help others.[1] Certainly one of the reasons that many large corporations are “good citizens”. They realize that if they don’t get involved and make alliances, there is a greater chance that they will be viewed to the negative and laws will be passed and public sentiment will make their business model less profitable.
I think that the problems of prosperity are the best kinds of problems to have. In a world where you are so innovative and valuable that you are paying software developers over $100k a year and everyone wants to live where these companies are you can just throw money at the problem. Coase would certainly endorse this. This is why successful places like NYC and Boston also have high taxes. People are willing to pay higher wages for better teachers or a city analytics department or a public transit system.
The housing and real estate problem can be solved by changing zoning laws. In SF, there is no where to develop. In NYC, they have rent control. Many cities use zoning, regulations, landmarking and pure politics which causes imbalances in housing markets.
On education, volunteering is great. But, I think startups can have a meaningful impact by unbundling education. The way we educate kids is not suited to the way the future looks. We have an assembly line process in an artisan world.
We have an assembly line process in an artisan world.Assembly lines and factories, are, in theory cost efficient (at least in the current as opposed to the future). That why’s only Burger King “let you have it your own way” and the others choked on that (from memory, based on past advertising). [1]If someone is advocating for artisan, then expect costs to go up. Would it be better? Let’s say it would be. But how are you going to pay for it?In life (which is not fair as we know) sometimes musical chairs means someone will lose, simply because there is not enough money to pay for chairs for everyone.Of course I am one of the biggest critics of the fact that the entire education system needs changing, starting with everything flowing from SAT’s, US News and World Report, and legacy halos and filters. But unfortunately I don’t think it is actually possible, anymore than solving the spam problem. To much momentum and head start in the old system.[1] https://www.youtube.com/wat…
Isn’t volunteering and the 1-1 learning opportunities it provides a way to move beyond the assembly line? I’ve wondered for a while now why we can’t have a radically better digital mechanism to integrate community expertise into schools through cool workshops, field trips, PD, internships, etc etc etc
volunteering is great. on a one off basis it provides exposure and a taste to someone. but real education comes from consistently being with the person over time. might only be once or twice a week, but it has to be contact over time-and it can’t be passive. For example, a field trip to the art museum might kindle and interest in art-but to really pursue it they are going to have to take a class; get mentored etc over time.
“We have an assembly line process in an artisan world.” — pointsnfigures.I’m going to start citing JLM and you liberally away from AVC, with accreditation.Tech itself went through a transformation from the waterfall (assembly line process) to agile (artisanal, experimental, rapid iteration) frameworks.So it unbundled itself and can unbundle education too.
Would a VC firm donate 1% of their carry to a cause?
A big barrier is simply the cultural disconnect between the world of public policy pdfs, clerical bureaucracies etc and the world of disrupting through apps, growth hacking etc. There’s some cool developments like CfA, the US digital service, new city “chief innovation officers” etc though it’s still very much early days and frankly too much has focused on providing a digital face lift to antiquainted analog processesIs the biggest public problem we face today really that the DMV has an annoying UI/UX? The “civic tech” community likes to talk about how we have a delivery crisis. They’re correct in the facts. NYU’s GovLab estimates that only $1 out of every $100 in government spending is backed by evidence that the money is spent wisely.More than that, a confluence of trends challenge governments around the world to improve delivery: aging physical infrastructure, demographic pressures on pensions, the uncertain impact of climate change, educating the next generation in a rapidly changing world, and the economic reverberations of globalization and increased automation.What apps are going to “solve” that? To some in the open data movement, that may seem like too harsh a question. Yet the phrase “delivery crisis” gets invoked frequently at open data events and in fact the idea sparked the start of Code for America:“You need to pay attention to the local level because cities are in major crisis. Revenues are down, costs are up — if we don’t change how cities work, they’re going to fail.” [see wikipedia article on CfA]If that’s the mountain we face, shouldn’t we ask what it will take to really meet it? Code for America’s stated path to improving delivery has been to “show what’s possible.” And they’ve succeeded. They’ve clearly demonstrated that there is no magical barrier preventing governments from using cutting edge web tools.Yet as the delivery crisis illustrates, much more needs to be done. How might this movement mature from pioneering novel customer service improvements to the normal everyday work of government? How might the web enable new uses of public data that can drive the government delivery improvements that we need?Anyway I’ve obcessed over this question for a long time now and here’s the latest synthesis of that thinking if anyone’s curious: argolabs.org/manifesto
NYU’s GovLab estimates that only $1 out of every $100 in government spending is backed by evidence that the money is spent wisely.This is almost certainly true. But the trend towards evidence based policy typically doesn’t move the needle in the direction you want. Rather, Government starts with a policy and runs inquiries/committees/commissions to find the evidence to support it; now only $0.80 out of every $100 in government spending is backed by evidence that the money is spent wisely.
Im as much a information individual as anybody, but while the typical team think is that information is fresh enough to back up (fail to reject) a hypothethis, most information is loud, very subjective and abused. Moreover, observe what you ask for! once that information and its outcomes gets in the arms of a politician, look out!Spybubble
This was a very different discussion from the more traditional tech panels
It was the only topic I would agree to talk about. I’m so bored with all the regular topics
Would loved to have had a few more minutes of interaction and a deeper dive between you and Ron on the data privacy issue. Interesting dichotomy in perspective there.
How can you get bored of talking about unicorns? 🙂
Exactly
I think that’s a critique of the conference organizers more so than you having little else to speak about. And that’s disappointing and a poor prognosis for tech conferences to come.What if there were a tech conference where the speakers got to speak/converse about what they wanted to speak about? 🙂
Fascinating to watch the inversion of generations in this new age of the sharing economy, Snowden, and the NSA : Mr. Conway on the side of liberty, privacy, freedom to innovate, the young interviewer on the side of government intrusion & intervention. Certainly not the expected roles, very interesting.
I recently looked at a UK start-up used by Starbucks, & UK retail giant Marks & Spencer, to help employees volunteer their time to projects in their store’s neighborhood. In addition to time, the 2 companies recently donated $2M which staff use to support projects in their neighborhood . https://www.neighbourly.com/Fred's work teaching programming in schools is rad. As a Comp Sci grad with a 100 person start-up, I’d love to see schools using this platform, so I could see if there was appetite for us to teach Comp Sci & Business Studies in our home town, Bristol, UK.
Great point about the need for tech industry leaders to inform civic leaders about important issues. I believe civic leaders do care about these issues and want to be aware of and knowledgeable about them, but don’t know how to educate themselves to this end.