O'Reilly Radar Podcast - O'Reilly Media Podcast

Latest Episodes:

Richard Cook and David Woods on successful anomaly response Nov 03, 2016

Show notes

O'Reilly Radar Podcast: SNAFU Catchers, knowing how things work, and the proper response to system discrepancies.

In this week's episode, O'Reilly's Mac Slocum sits down with Richard Cook and David Woods. Cook is a physician, researcher, and educator, who is currently a research scientist in the Department of Integrated Systems Engineering at Ohio State University, and emeritus professor of health care systems safety at Sweden’s KTH. Woods also is a professor at Ohio State University and is leading the Initiative on Complexity in Natural, Social, and Engineered Systems, and he's the co-director of Ohio State University’s Cognitive Systems Engineering Laboratory. They chat about SNAFU Catchers; anomaly response; and the importance of not only understanding how things fail, but how things normally work.

Here are a few highlights:

Catching situations abnormal

Cook:

We're trying to understand how Internet-facing businesses manage to handle all the various problems, difficulties, and opportunities that come along. Our goal is to understand how to support people in that kind of work. It's a fast changing world, mostly that appears on the surface to be smoothly functioning, but in fact, as people who work in the industry know, is always struggling with different kinds of breakdowns, and things that don't work correctly, and obstacles that have to be addressed. Snafu Catchers refers to the idea that people are constantly working to collect, and respond to, all the different kinds of things that foul up the system, and that that's the normal situation, not the abnormal one.

Woods:

[SNAFU] is a coinage from the grunts in World War II on our side, on the winning side. Situation normal, so the normal situation is all fucked up, right? That the pristine, smooth work is designed, follow the plan, put in automation, everything is great, isn't really the way things work in the real world. It appears that way from a distance, but on the ground, there are gaps, uncertainties, conflicts, and trade offs. Those are normal—in fact, they're essential. They're part of this universe and the way things work. What that means is, there is often a breakdown, a limit in terms of how much adaptive capability is built into the system, and we have to add to that. Because surprise will happen, exceptions will happen, anomalies will happen. Where does that extra capacity to adapt to surprise come from? That's what we're trying to understand, and focus on, not the SNAFU—that's just normal. We're focusing on the catching: what are the processes, abilities, and capabilities of the teams, groups, and organizational practices that help you catch SNAFU's. That's about the anticipation and preparation, so you can respond quickly and directly when the surprise occurs.

Know how things work, not just how they fail

Cook:

There's an old surgical saying that, 'Good results come from experience, and experience comes from bad results.' That's probably true in this industry as well. We learn from experience by having difficulties and solving those sorts of problems. We live in an environment in which people are doing this as apprenticeships very early on in their life, and the apprenticeship gives them opportunities to experience different kinds of failure. Having those experiences tells them something about the kinds of activities that they should perform, once they sense a failure is occurring. Also, some of the different kinds of things that they can do to respond to different kinds of failures. Most of what happens in this, is a combination of understanding how the system is working, and understanding what's going on that suggests that it's not working in the right sort of way. You need two kinds of knowledge to be able to do this. Not just knowledge of how things fail, but also knowledge of how things normally work.

No anomaly is too small to ignore

Woods:

I noticed that what's interesting is, you have to have a pretty good model of how it's supposed to work. Then you start getting suspicious. Things don't quite seem right. These are the early signals, sometimes called weak signals. These are easy to discount away. One of the things you see, and this happened in [NASA] mission control, for example, in its heyday, all discrepancies were anomalies until proven otherwise. That was the cultural ethos of mission control. When you lose that, you see people discounting, 'Oh, that discrepancy isn't going to really matter. I've got to get this other stuff done,' or, 'If I foul it up, some other things will start happening.'

What we see in successful anomaly response is this early ability to notice something starting to go wrong, and it is not definitive, right? If it was definitive, then it would cross some threshold, it would activate some response, it would pull other resources in to deal with it, because you don't want it to get out of control. The preparation for, and success at, handling these things is to get started early. The failure mode is, you're slow and stale—you let it cook too long before you start to react. You can be slow and stale, and the cascade can get away from you, you lose control. When teams or organizations are effective at this, they notice things are slightly out, and then pursue it. Dig a little deeper, follow up, test it, bring some other people to bare with different or complimentary expertise. The don't give up real quick and say, 'That discrepancy is just noise and can be ignored.' Now, most of the time, those discrepancies might probably be noise, right? Isn't worth the effort. But sometimes those are the beginnings of something that's going to threaten to cascade out of control.

Sam Wang on predicting the election, finding truth in mass media, and a new role for the cerebellum Oct 20, 2016

Show notes

The O'Reilly Radar Podcast: Prediction algorithms, cognitive biases, and how our brains come online.

On this week's episode, I chat with Sam Wang, professor of neuroscience and molecular biology at Princeton. Wang is also a co-founder of the Princeton Election Consortium, a site focused on analyzing and predicting U.S. national elections. We talk about the site's prediction algorithm and this crazy election cycle, and the role neuroscience may have played. We also talk about the current research Wang and his team are working on, the U.S. BRAIN Initiative, and the powerful role governments play in academic research.

Here are some highlights:

Predicting elections

What we do at the [Princeton Election Consortium] site is we collect publicly available polls. This is a website that's at election.princeton.edu. We take those polls and then feed them into a script that I've written and scripts that my students have written, and on an automated basis, we take those polling data and we turn the polling data into a clear, sharp snapshot of exactly where the presidential race appears to be at any given moment, on any day during the campaign. It's sort of a tracking index that says what would happen on election today, and we do it using what I would call optimal statistical tools.

... The Princeton Election Consortium is open source, and so anybody can download the scripts. They're written in MATLAB and in Python, and there are some shell scripts. Anybody can download the stuff and run it for themselves. ... We take all the available state polls for a given state, say Virginia, for example, which is a competitive state, and we take the median margin for that state between Hillary Clinton and Donald Trump, the median margin of all the polls, and that gives the best estimate of what the likely margin is going to be in the election. Then we use the spread in that set of data to figure out how probable it is that Hillary Clinton or Donald Trump is in the lead.

That's one probability. Then we do that over and over again for all 51 races, the 50 states plus the District of Columbia. We do it over and over again, and in each case there's an outcome that's like a coin toss, except that the coin toss is worth some number of electoral votes, and then we combine all those probabilities using just some simple math trick, a function in MATLAB that's called convolve.

We turn all those probabilities into an exact distribution that has a lot of sharp peaks in it corresponding to particular combinations of states. It's anywhere from zero to 538 electoral votes for Hillary Clinton and the same for Donald Trump. All that's done automatically, and then that tells us how conditions are today. Then on top of that, we add some assumptions about where things are likely to go by Election Day, and that's a random drift factor. That random drift factor gives us a view of what is likely to happen on November 8.

Getting to the truth

Getting back to the general subject of mass media and also narrowcast media, we have all kinds of sources of information available to us now, Facebook, Twitter, news feeds, talk radio, e-mail. We have all these channels of information, and it's really super hard for any individual to really cut through all that clutter and get accurate information, even if some of those channels of information are high quality. I think that actually these cognitive biases make it super tough for citizens to really get by in what should be a golden age of information.

... Another big challenge that not any one person can address is figuring out how to create a media ecosystem in which right information is more likely to get into people's heads. For example, it's actually not a bad thing, for instance, for media organizations to bring on people of opposing viewpoints, because bringing in people of opposing viewpoints actually causes ideas to get examined critically, and it's possible to have everything in one neat package where you get to see it all at once. Just to give you an example, it actually gives people a direct comparison between Clinton and Trump to see them next to each other. Debates have an important function where you see a direct interaction between these two very different candidates. Anyway, getting back to getting to the truth of things, I think that a major challenge for media is how to get information across accurately to people and have it stick.

Finding the development source of coordinated thought

In the lab, we're super excited about some stuff we're doing now trying to understand how cognitive and social abilities arise in the brain. For the last 15 years at Princeton, I have been interested in how the brain changes in response to experience, so how learning happens and how development happens. For a lot of that time, I've been interested in sort of the nuts and bolts of how single connections work. We use optical methods and advanced molecular biological methods to watch brain activity in action. We can do things like use optical methods to watch a brain circuit in the brain of a mouse as the mouse is navigating a virtual maze or as the mouse is investigating some puzzle that it has to solve, like a simple maze.

A lot of what we've been doing is just understanding how the brain integrates information to learn from its environment. That involves multiple brain systems, and I study a brain region called the cerebellum, which, if you look in textbooks, it's usually a brain region that's important for balance or movement, but the way it does that is by integrating sensory information to try to keep mental processes on track.

In the last few years, we've become very interested in the possibility that the cerebellum controls not only fine actions in adult life, but it might even act to shape the growing brain. One piece of information we have that suggests that is some clinical evidence that others have noticed, which is that if babies by accident have some injury to the cerebellum at birth, so if there's a difficult birth like a bleed, then the odds of autism go up by a factor of 40. It's bigger than the cancer risk that comes from smoking.

What we suspect, based on that, is that maybe the cerebellum is some kind of guide that plays an absolutely necessary function for babies to grow their mental capacities. What we think is just as the cerebellum guides coordinated movement, what we think is that maybe the cerebellum acts to guide the development of coordinated thought, where babies learn to recognize faces and voices and pick up language. There are all these incredible things that babies do. We're really deeply immersed in testing the idea that this is a part of the brain that helps teach the rest of the brain to come online.

John Bassett III on the global economy, the power of people, and how to make it in America Oct 06, 2016

Show notes

The O'Reilly Radar Podcast: Navigating the increasing globalization of industry and commerce.

In this episode of the Radar Podcast, I chat with John Bassett III, chairman of the board of the Vaughan-Bassett Furniture Company. We talk about globalization and the effect it's had on the furniture industry, the international trade battle he waged (which was written about by Beth Macy in her book Factory Man), Bassett's book Making it in America, and what entrepreneurs need to know to succeed in business today.

Here are some highlights:

Panic sets in

They started making furniture in China, and we competed very well through the 1990s. All of this changed dramatically in 2001 when they became a member of the World Trade Organization (WTO). Once they were in the WTO, their prices plummeted, and the bottom dropped out of the market. By that time, I had left the Bassett Industries and joined my wife's family furniture company called Vaughan-Bassett Furniture on January 1, 1983. I was over here at the time, but the whole furniture industry in wood was affected. Factories were closing left and right. Thousands of people were being laid off. There was panic. That's the only way to explain it.

Then we found out that there was a rule at the WTO and a law on the United States books. The law actually goes back to the 1930s, it's called dumping. Dumping is when you sell a product in another country for less than your manufactured cost, and what you're doing is dumping your product in that country to force everybody out of business so you can capture all of the business. That's exactly what happened. I led a coalition that challenged the Chinese, and at that time, it was the largest dumping petition brought against the government of China ever at the WTO.

Going to battle

Prices were plummeting, and many of the United States manufacturers at the time said, "Well, we'll close our factories and just buy this product overseas. That's what we'll do." I wanted to go and actually look the people in the eye. I wanted to see exactly what was going on. I went to China and I went to Northern China. The prices seemed to be less up there than anywhere else.

I met this gentleman who was erecting a huge, well, a series of factories with obvious Chinese government help. I told him, 'I might be interested in buying your product.' He looked me in the eye and said, 'This is what you must do.' I said, 'All right.' He said, 'The first thing is you must close every factory you have. You must get rid of all of your people. You must sell all of your machinery and you must put yourself in my hands.'

There was no smile on the face. He was extremely serious. As we left to fly back to the United States, I told my son Wyatt, 'Get ready. We're going to war. They are being supported by the Chinese government. They are picking up the bill. This is not what we were promised when the Commerce Department asked us to support GATT. They're dumping, and either we're going to have to resist this or this industry will disappear.'

Your power is in your people

I knew then the rules of the game had changed. ... We just had to adjust the way we ran our businesses. Everybody talks about innovation, education, entrepreneurship, all of that, and I agree with all of it. We did something different. I wrote a book after Beth [Macy] wrote her book, Factory Man. I wrote a book called Making It In America. We organized our people and our organizations. Before we shut everything down, and we did close some factories, but we went to our people and we said, 'If we're going to survive, we've got to do this together.' The book is about how we organized our people.

The people in these plants wanted to be a part of this. They did not want some CFO looking at figures and closing the plant. They said, 'We can make a better product. We can make a less expensive product, and we can deliver it faster and we can do all these other things.' The American worker is an exceedingly efficient worker, but you have to give them a chance.

Playing by the rules

My position is this. There are rules of the game out there for everybody in the WTO, including the Chinese and the Indians and others. Let's play by the rules. Donald Trump talks about new laws. We don't need new laws. We need to enforce the laws that we've already pledged to do. Let me give you an example. In the anti-dumping law, I went back to when we started our petition, which was 2003 through 2015. I took three countries: China and India, the two that certainly have the largest population and probably the most to gain, then I took the United States, which probably has the most to defend, being the largest market. I looked at how many dumping petitions have been imposed, not initiated but actually imposed, by these countries against other countries over that 11- or 12-year period. India leads the list. They imposed 353 anti-dumping petitions against other members of the WTO. Number two was China at 166. Number three was the United States at 163. The country that had the most to defend imposed the least. ... I think there are many benefits to globalization, but when countries cheat, they should be called to task for it.

How to make it in America

I would offer new entrepreneurs several pieces of advice. Number one is, if you're going to play on this ball field and if you're going to play in this game, be sure you're adequately capitalized. A lot of the people that you're going to compete against have staying power, so be sure you have enough capital to take on whoever your adversaries are going to be. Number two is, don't overlook the power of your people working for you. And, obey five of the 12 rules [I outline in my book], what I call the Five Great Rules: number one, attitude. You have to start with an attitude of 'we're going to win.' Don't start as a loser. Two: leadership. Don't ask anybody to do something you won't do yourself. Roll up your sleeves and go to work with your people.

Three: Change. When you start out, be willing to change because things move so fast today, what you do today might not be relevant six months from now. It might not be relevant six weeks from now. Number four is, don't panic. They love to panic you and tell you you can't do it. The easiest battle to win is when the other side surrenders before the first shot is fired. Just calm down. There's never been a good business decision made when people were panicking. Last is teamwork and communications. Everybody in your organization has to be on board, and the way you get them on board is through communication. Constantly tell people where you are and ask for their help. Those would be the things I would tell a young entrepreneur to do.

Haakon Faste on designing for a "post-human" world Sep 22, 2016

Show notes

The O'Reilly Radar Podcast: perceptual robotics, post-evolutionary humans, and designing our future with intent.

In this Radar Podcast episode, I chat with Haakon Faste, a design educator and innovation consultant. We talk about his interesting career path, including his perceptual robotics work, his teaching approaches, and his mission with the Ralf A. Faste Foundation. We also talk about navigating our way to a "post-human" world and the importance of designing to make the world a more human-centered place.

Here are a few highlights:

Multimodal interface systems

What these robotics systems allow you to do, which is really exciting, is you can take someone who's an expert at a certain scale, maybe they're an athlete, and you can put them into a robotic system and ask them to perform their craft, if you will. You can imagine taking an expert, someone like Tiger Woods who's a fantastic golfer, having him climb into a robot suit and show his perfect golf swing, record it, and then have novices climb into the robot suit, hit play, and sort of play back the expert's body knowledge into the novice's body.

This is the notion of a multimodal interface. You can couple all of the modalities of sensing: the visual sensation of being in a situation, the sound effects, and then haptic feedback, whether that's force feedback on your gross body movement or specifically simulating what they call pseudo-haptics, or the sensation of touch. You can put little vibrating motors all over your body and make a responsive suit.

We were interested in studying what happens if you put someone in one of these systems and show them visually how it should be, and then ask them to perform, or you play it back into their body and ask them to perform it. You can study how quickly people learn how to do those skills. This is a really powerful set of technologies for things like post-stroke rehabilitation, areas where people have lost their ability to use their body, and other kinds of situations. You can imagine if you try to teach a robot how to walk or how to perform a skill, it's very important that you be able to capture the skill in the first place so that the machine can learn from you.

Designing intuitive experiences

First, we spend a lot of time learning how to use our body, and then we move into a kind of visual stage, and then during adolescence we learn to deal with our emotions, and finally learn skills of higher order, reasoning, and critical thinking, and symbolic thinking, and so forth. Typically, what works well from a perceptual standpoint is just to recognize that a lot of the things that we consider to be expertise in our use of technology are thinking about it from this adult symbolic perspective.

We presume that if we tell someone something with words, they will do what we tell them to, whereas in actuality, they're going to respond very automatically from an emotional level and even at a lower-level embodied experience of the world. From a perceptual perspective, those systems are deep in our control system as humans, and they're very reflexive and automatic, and that's the source of our intuition.

So, when we're trying to design experiences that make things, say, more intuitive for someone, it's really important that you leverage the aesthetic and feeling-based and emotional aspects of an experience because it more immediately connects with what is intuitive to them. Of course, you never know how an experience will be used until you observe people using it, and a lot of times your hunches are quite wrong, which is why designers use methodologies around rapid prototyping and iterative design to get stuff quickly into the world without presuming that we know what's going to work.

Giving our tools the capacity to shape our future

Humans have evolved to have a certain set of capacities when we interact with the world. Today, we live in an experience that has all of these new technologies that are really shaping the way that we think and act. We have computers, and we have the Internet, mobile devices, augmented reality, and other things that fundamentally alter our sense of what a human is—so, an organ transplant, or a drug that is designed to have some kind of emotional effect, or genetic engineering. These are capacities that are fundamentally changing the biological nature of what humans are.

The post humanist theorists called this a kind of trans-human state or transitional human, and what we're moving toward is this kind of hypothetical—it's hard to pin down what the future will be, of course, and we don't want to be overly deterministic—but I think we can say pretty confidently that we're going to have capacities that radically exceed those of our present biological capability. The theory goes that then we sort of transcend the unambiguous nature of what it means to be human and we become something else.

When you mix into that super intelligence or autonomous robotics or life extension—the ability to live forever, that you could encode your mind and upload it into the cloud, or design you own children, or simulate possible variants of yourself, try a variety of different medical treatments and then pick the one which survives—we're moving out of a state of humans as designed by nature, evolution, through this contemporary trans-human state of a world that's designed by humans because we design our media and our technology and so forth into a state where we are no longer human because we're sort of post evolutionary. We've given our tools the capacity to shape our own future.

Designing the future with intent

You have to recognize as a designer you are always doing what you're doing in the service of power, and by doing that project you're perpetuating the values of whatever it is that's driving that broader intent.

As a designer in that system, you need to be very cognizant of your own values, what you are and are not willing to do. You need to push back when you feel like there needs to be pushback. We live in a very subjective world, and it's incredibly complicated, and everything is double sided. You need to be comfortable for yourself about what you're doing, but I really think it's important that you have a much bigger sense of what the world needs and that you are working toward those things that the world needs.

Because we're entering a world, hopefully, that's increasingly democratic and distributed, and that values diversity, and values different perspectives, and creates services and all of these different little niches that benefit people in all kinds of nuanced, magical ways. It's very important that you keep those systems open and that you have a strong point of view about the kind of future you want because it would be so easy for very powerful lobbies, politically or technologically if you will, to own all of the data or all of the thinking and have everyone else kind of follow like sheep.

Pete Skomoroch on the current state and future potential of bots Sep 08, 2016

Show notes

The O'Reilly Radar Podcast: Bot hype, bot UX, and bots in the workplace.

This week on the Radar Podcast, we're featuring the first episode of the newly launched O'Reilly Bots Podcast, which you can find on Stitcher, iTunes, SoundCloud and RSS. O'Reilly's Jon Bruner is joined by Pete Skomoroch, the co-founder and CEO of Skipflag, to talk about bots—about what's driving the sudden interest, what we can expect from the technology, and some interesting emerging applications.

Here are some highlights:

The uncanny bot valley

I've seen a lot of hype waves over the years in tech, but this one is growing pretty rapidly. That exact story I've heard from a few people, where CIOs from big companies are actually saying, 'All right, what's our bot strategy? I want to stop, I want to retask some people to dig into this.'

There's been a lot of things like that in the past, where it could feel misguided because, 'Wait, it's too early—we don't even know what this is yet.' At the same time, there's usually something behind these things. Another recent analogy was Minority Report, right? If you go back to 2002 when that movie came out, the boardrooms were echoing with, 'I want an interface like that! I want to talk to a computer with my hands and wave them around.'

Now, maybe a little bit of what we're seeing is like the movie Her, which came out in 2013. ... It's kind of eerily close to where we are, it feels like, but there is that uncanny valley between what you see in the movie and where the AI tech is right now. I think that's why it feels a little bit like hype—most people don't grasp the difference.

1,000 bots versus one god bot

Benedict Evans at Andreessen Horowitz has been writing a bunch on both the rise of messaging over the last four or five years, and now he's talking a lot more about conversational commerce and UX and bots. I really liked one quote he had, which was, 'What can I ask if I can't ask anything?' This is a different kind of discovery, right?

Before, we were talking about discovery of apps, discovery of bots or products. There is a deeper problem, which is, when I'm in a conversation with a new bot, if the interface for every bot is kind of the same, it's some text interface, it's unclear exactly who I'm talking to and what they know and what they don't know and what I can ask. If it has some knowledge inside the bot's memory, it's unclear what it knows and what it doesn't know.

That's where I think Amazon Alexa—they're walking a line, but I think part of the reason it's clicking with some consumers better than previous attempts at these things is, my understanding is, they spent thousands and thousands of hours with actual voice actors in a room asking it a lot of different questions, and then, kind of brute force training it to respond well and be resilient to these kinds of requests.

Now, that's not a realistic solution for most other bots, and I think part of the solution here is going to be either better UX in these messenger platforms, so that you could have a more clear sense of the options and of the menus, if you are texting. Then another thing is being very clear about what the bot is good for and what it isn't. This is more like 1,000 bots versus one god bot.

Overcoming the brittleness issues of the semantic web

If you go back to the semantic web days, the vision was that you'd have this machine-understandable interface so that machines could talk to machines, and all these queries, like booking a flight, would magically happen. The vision that everybody really wanted was—Apple had this vision of the Knowledge Navigator.

We're actually, I think, not that far off from that demo these days, but it's kind of a walled garden demo, where you could build that for that specific case, but to enable almost any generic application, what you really need is a fuzzy way for APIs to talk to APIs with some reasoning and intelligence. I don't know if this bot wave is going to stick or if your bot strategy is going to really matter at the end of the day, but I'm actually optimistic that machine learning is going to keep cranking away.

Text is here to stay; it's a nice way to talk to people in public without everybody talking over each other. What is interesting is we're training machines now to talk via text. Now, what happens when you have a machine talk to another machine via text? Do we get over some of those brittleness issues that killed things like the semantic web?

Bots at work

I'm pretty bullish on the idea of AI in the workplace. That's why I'm pretty excited about the Slack platform. They were one of the early movers. Once they called the apps that you could build on Slack 'bots,' I think that's really where you saw a step function in the number of bots, because by definition, if you're building an app on Slack, it's a bot. Now, Facebook has followed suit, and everything there is a bot as well. I think you're going to see this split between e-commerce applications, and then in the workplace, I'm sure a lot of the big workplace players will have some form of bot platform or bot interaction.

Cory Doctorow on nascent pro-security industries Aug 25, 2016

Show notes

In this O’Reilly Radar Podcast: The impact of minimal IoT product security and the case for new pro-security business models.

This week's Radar Podcast episode is a special cross-over edition from the O'Reilly Security Podcast, which you can find on iTunes, Stitcher, RSS, or SoundCloud. O'Reilly strategic content director Courtney Nash chats with Cory Doctorow, a journalist, activist and science fiction writer. They talk about nascent pro-security industries, the EFF's lawsuit against the U.S. government, and the new W3C DRM specification.

Here are some highlights:

Auditing IoT products is a liability for security researchers

Think about the conditions under which IoT companies operate. Their business plan—the thing they show to VCs to get the money to go into the business—is to monetize data. They're all designed with security as an afterthought. They're all designed with the minimum viable security to make this product not immediately burst into flames after you put it inside your body or put your body inside of it. Even worse, security researchers face total, brutal liability for investigating these devices and telling people which ones are and aren't safe. It is completely nightmarish.

New pro-security business models

Note: The Electronic Frontier Foundation is representing Bunny Huang and Matthew Green in a case challenging the constitutionality of Section 1201 of the DMCA.

One of the things that our DMCA lawsuit would provide for is a pro-security business model. Imagine if you could start a commercial consultancy that would come in and deworm your IoT household. It could come in and jailbreak all the devices and check their firmware loads, and replace the firmware loads with open firmware or patched firmware, or something else that sits in between. All of those things, all that commercial stuff as well, is currently off-limits, and would be available in the same way that you can enable third-party parts and services if there are no legal impediments. The hardware service and support market in the U.S. for all classes of goods, from lawnmowers to cars to air conditioners to computers, is 2 to 4% of America's GDP. It's a gigantic multi-billion-dollar sector, and in many cases, these are small and medium-size enterprises.

Related resources:

The EFF is suing the US government to invalidate the DMCA's DRM provisions (BoingBoing)
America's broken digital copyright law is about to be challenged in court (The Guardian)
The 1201 complaint in full

Alyona Medelyan on applications of NLU Aug 11, 2016

Show notes

The O'Reilly Radar Podcast: Natural language understanding and natural language processing applications, our future with chatbots, and open source indexing.

This week, I talk with Alyona Medelyan, co-founder and CEO at Thematic and founder and CEO at Entopix. We talk about natural language understanding, the challenges of analyzing unstructured text, and her open source indexing tool Maui that she's been working on for the past 10 years.

Here are some highlights:

Use cases of Natural Language Understanding

Natural Language Understanding is really a sub area of Natural Language Processing (NLP). In general, NLP deals with using computers to understand human language, but not all NLP tasks require actual understanding. For example, if we take part of speech tagging, when an algorithm decides whether a word is a noun or an adjective or a verb, in order for the the algorithm to perform this accurately, we don't really need to know what the words mean. You can achieve quite a lot by simply counting how many times part of speech text follow each other, and very simple techniques would be sufficient. On the other hand, if we're building a dialogue agent, a chat bot like Siri for example, in order to respond meaningfully, Siri would need to understand what each of our statements mean, and this is where the understanding comes in.

Practical applications of NLU for enterprise

A lot of what can be done with NLU is very practical. I'm actually in Portugal at the moment, and I don't know any Portuguese. Every time I go to a restaurant or buy groceries or search for places, I use Google Translate, so it's quite practical. In terms of what everyday businesses, not just giants like Google and Apple, can do with NLU, I think the key example would be understanding customer feedback because these days, pretty much everybody has a smart phone. Everybody has written review for a company if they like their services or they didn't. People will send complaints and so on. With all of this text, businesses become more competitive because they know people can read all these data. Sentiment analysis—one of the techniques that uses natural language understanding to not just understand whether the customer is happy or sad, but also what are the specific things they're saying the business is good at or which ones they can improve—this can practically help them to compete and get better at their offerings.

Maui: More than a digital librarian

In a traditional library, a librarian categorizes books so that people can find them. In a digital library, Maui takes this role identifying what each book or each document is about. This is what Maui does; its results can be used to improve search and organize documents, but that's just one of the applications. I also helped companies apply Maui in many interesting ways. One company used it to link advertisers to web pages to display content-relevant ads. Another used it to send users content recommendations. How it differs from Thematic, is Thematic is specially designed to analyze short pieces of text, something that Maui doesn't do well. Maui works great on written documents where people actually thought about how to write them, and Thematic works better on short text and can detect more fluid themes than Maui.

Our future with chatbots

I think that chatbots and automated personal assistants, even though currently are not particularly well advanced in what they're doing and require a lot of humans helping, will still become more prevalent in the future. That would mean that we won't need to interact with people as often. Just like online banking made the cost of making transactions cheaper, customer support will become cheaper, too, thanks to chatbots.

On the other hand, businesses will compete on providing the best deals and the best customer service for their customers. I think they will use more and more natural language understanding to figure out what people say about their business, about the competitors, about the products. In the end, we as customers will be the one who will benefit from all of this.

Designing better security outcomes for human beings Jul 28, 2016

Show notes

The O'Reilly Radar Podcast: Eleanor Saitta on security countermeasures at the human level, the relationship between security and design, and understanding security design as a separate discipline.

This week's episode features a special cross-over conversation from the O'Reilly Security Podcast, which you can find on Stitcher, iTunes, SoundCloud, or RSS. O'Reilly's Courtney Nash chats with Eleanor Saitta, a security architect at Etsy. They talk about the importance of thinking of security in a human context and the increasingly critical relationship between security and design.

Here are a few highlights:

Detecting fraudulant patterns at the human level

Look at banking fraud and fraud detection systems. Although financial malware is a real issue, and we are seeing more and more people who end up with malware running on their phones that then attacks bank authenticators or logs into their account and makes transfers. These are starting to be very real issues, let alone credit card numbers and all this kind of stuff. The biggest way that those attacks are stopped isn't by preventing code from running on people's machines, it's by detecting fraudulent patterns and transfers at the human level, and cutting things out at business rule levels, and much higher levels.

In the worst case, it's someone goes into a bank physically and talks to someone, and has a conversation. That's just as much a part of the security countermeasure set as any number of anti-banking Trojan, anti-malware projects are.

The relationship between security and design

That whole process of coming into understanding the high risk world a little bit more was really, in some ways, it was really challenging for me because I'd spent probably eight years, nine years at that point when I first started getting involved in that community, doing big enterprise security. To come into this community and to realize that actually I know very little about how to create better security outcomes for human beings was an interesting thing to learn midway through my career.

What it made me do was go back and think a lot about the relationship between security and design, and realize that one of the things that we need to do when we're building systems for, at the time, I was mostly thinking about high-risk people, but I've realized that this applies to any system. We need to understand not just what that user is worried about, but what the countermeasures that they can use to cancel out their adversaries attacks are, because we're dealing with that design space much more than we are with the code space. Now, if we can find things at the code level that give us new capabilities in that design space, that's amazing. So, being able to get rid of classes of low-level bugs, so we can stop thinking about them—great, that's a huge capability for the design space and the architecture space. All of the different things that we can do with cryptography, as far as using it to reduce the kinds of attacks that people can be subject to and giving them new invariants the system can let them use. Great, amazing capabilities, but the reason why they're interesting is because of how they shift that design space, and that has to be the thing that starts driving everything.

Security design as a separate discipline

There's a conversation between architecture and requirements and design. There has to be. None of these can act independently, but the thing that we don't see, the thing that I really don't see in the security community yet, is an understanding of security design as really a separate discipline. This is literally what I'm spending my time doing right now.

Othman Laraki on achieving the long-tail distribution of genetic insights Jul 14, 2016

Show notes

The O'Reilly Radar Podcast: Color Genomics, genetic testing access, and the future of precision medicine.

This week, I chat with Othman Laraki, co-founder of Color Genomics. We chat about challenges and opportunities in genetic testing, the future of precision medicine, and the hurdles medicine and health care are currently facing (and how we can overcome them).

Here are some highlights:

Genetics testing for everyone

Genetics, we felt, had come to a point where there was an opportunity to have a very big impact by essentially mixing some of the best of the biology world with software—in many ways, genetics had started to become, in part, a software problem.

It felt like it was starting to be possible to build products that made genetics accessible to a much broader population by both dropping costs as well as increasing access, so making this information more accessible to a much broader population in a scalable way.

... For example, one of the things we did that we're very proud of is we created this program called the Every Woman Program, where whenever someone buys a test from Color, they can also contribute to fund testing for someone who can't afford it. Then we work with a number of cancer centers, for example at UCSF and the University of Washington, Morehouse in Georgia, and a number of others, where each one of those centers works with underprivileged populations, and they can provide tests for free for people who can't afford it but who the doctors think should get tested.

Opportunities in machine learning

One of the big opportunities for machine learning in genetics, for example, is around the interpretation of the effects of specific genetic changes. Right now, there are set of guidelines or processes that are used by the industry around the interpretation of how a specific mutation impacts a gene. It's a structured process that's very labor intensive, but it's one of those areas where over time is going to become something that's very heavily solved by machine learning because there's a lot of data that can be used to train a model instead of purely running it in a manual way. The industry is going to evolve quite a bit over the next few years and machine learning is going to have very substantial impact there.

Using the full data set of the human body

Each one of us is carrying and generating a tremendous amount of data in our daily lives, whether it's our genome, our microbiome, etc., etc. So far, the link between that data and health practice had been through the path of research and translation to a few proxies, essentially, where researchers collect a lot of data, they do a research study, it turns into a set of conclusions, and that over time gets turned into a few rules that get introduced into medical practice. If someone's lipid levels are at this level, etc., then you draw these kinds of conclusions.

Now, we're coming to the point where the amount of data that a doctor will be able to use in a real way to make medical decisions is going to be the full data set of our bodies, which is very exciting and can have a very big impact.

Long-tail distribution of genetic insights

In some ways, I feel right now we've come to this point where there's been enough data and science behind us that we can already create a lot of value, and that allows the bootstrapping of doing things at a massive scale that really takes us to that long-tail distribution of insights around how genetics work and how the body works.

Smart cities need to inject a dose of humanity to be truly great cities Jun 30, 2016

Show notes

The O'Reilly Radar Podcast: Conversations with Daniele Quercia and Frank Cuypers.

This week's episode features two conversations I've had recently centered around smart cities. First, I chat with Daniele Quercia, research team lead at Bell Labs. We talk about research he's working on now; the launch of goodcitylife.org (including smelly maps and happy maps); why our use of technology shouldn't just aim to make a city smart, but to improve the day-to-day quality of life of it's citizens; and about the emerging areas of urban informatics he's finding most compelling.

In our second segment, I chat with Frank Cuypers, associate professor at the University of Antwerp and strategist at Destination Think! We chat about the importance of urban DNA, his nonprofit project Why Your City, and why there's no such thing as a smart city.

Here are some highlights:

Alternative smart city agendas

Daniele Quercia: "The idea and the rhetoric behind Smart City is one of efficiency and security. Usually they say a smart city is a city in which, if you go to work, you're always going to be on time. If you go shopping, there is no queue. You know what? You're going to feel really, really safe because of the CCTV cameras around you. It's about efficiency, security. We all know that we don't choose a city because of just efficiency and security. They make a city acceptable, but they don't make a city great. What makes a city great are fuzzy concepts, concepts that are really difficult to quantify, like beauty, happiness, these sort of things. That's why we built goodcitylife.org, which is a global network of people—researchers, people in industry, who really think there is an alternative agenda to the smart city agenda, and that's what we want to do. We want to empower these people, and we want to also do research in this new area of simply giving a good life to people."

Daniele Quercia: "Currently, I'm thinking about something related to algorithmic regulation. We wrote a paper that we presented a few weeks ago at the conference Dub Dub Dub Dub. The idea was that, basically, these platforms like Airbnb or Uber generate data. These data could be used for regulation. What we found out is that, for example, we look at the evolution of all Airbnb in London, and we look at which areas were affected. For those areas, we had census data. The idea was that now you can see how the evolution of Airbnb is related to different social economic conditions. Then you can see that certain areas, there might be some dodgy subletting going on, and you can regulate that because you can build an index after the data. The same thing, in general, that you can do with any platform, right? You generate data inside a city and then you take that data to build analytics that might be useful for policy-making? In theory. You can change your policies and then you can see the impact of those policies in the city, and so on."

"Cities are for humankind what the telephone booth is for Clark Kent"

Frank Cuypers: "Of course, I'm a fan of smart cities. The thing is that ... Let's talk about data. Seth Godin wrote, 'Data gets us the Kardashian's.' People become lazy. There are two many stakeholders that are sharing data, data, data, and they don't have a clue what they are talking about. I have never met a politician who knows the difference between long data, big data, and short data. What is worrisome is that we're building now sort of vendor-lock in places. For instance, New York. The council writes a letter to Google to ask: in Google Maps, it's always that with your car, you go to the left. That causes a lot of traffic jams. If 30% in their maps could go to the right, it would solve the problem. They didn't answer, and I don't know whether they have answered already, but that's not the point. Is it normal that you have to ask a private company to solve a traffic problem in a public space?

"What we need actually is—Jane Jacobs again, grassroots activism online. I say there is no such thing as a smart city because I really believe that cities are for humankind what the telephone booth is for Clark Kent. He goes in and in a split of a second, he transforms into Superman. We know that cities are the places where we could always transform our technology. It's about technological disruption as well, and that's a very good thing. But bring these two big trends together—on the one hand, you have the GNR, the genetics, the nanotechnology, the robotics, information technology. On the other hand, urbanization. 80% will live in cities in 2060. This is the most important thing we have to fix in our lives."

Frank Cuypers:"Now, I'm very interested in people who go to the basic question, what is the economic foundation of our city? Someone who I really appreciate, and he's not so known in the Anglo-Saxon world but he's advising the Vatican and the State of Ecuador, is Michel Bauwens with his peer-to peer movement, where you really, really have a sharing economy. Uber and Facebook are in my world, but in my vision, they are not part of a sharing economy. They create a win-win situation: there is something free for you and there is some money to earn with your data. In the end, there is someone who always pays the bill.

"It's the same with tourism. Sometimes we create a win-win situation, but the environment is destroyed, like the Great Barrier Reef in Australia. We need to create a win-win-win situation, without any party losing anything. That kind of disruption in economy and in politics I think is very necessary in these days because we can't keep pace with what happens in technology. We can't keep pace with what happens in information technology."

Our TOPPODCAST Picks

Follow Us

Stay Connected

O’Reilly Radar Podcast – O’Reilly Media Podcast

Catching situations abnormal

Know how things work, not just how they fail

No anomaly is too small to ignore

Predicting elections

Getting to the truth

Finding the development source of coordinated thought

Panic sets in

Going to battle

Your power is in your people

Playing by the rules

How to make it in America

Multimodal interface systems

Designing intuitive experiences

Giving our tools the capacity to shape our future

Designing the future with intent

The uncanny bot valley

1,000 bots versus one god bot

Overcoming the brittleness issues of the semantic web

Bots at work

Auditing IoT products is a liability for security researchers

New pro-security business models

Related resources:

Use cases of Natural Language Understanding

Practical applications of NLU for enterprise

Maui: More than a digital librarian

Our future with chatbots

Detecting fraudulant patterns at the human level

The relationship between security and design

Security design as a separate discipline

Genetics testing for everyone

Opportunities in machine learning

Using the full data set of the human body

Long-tail distribution of genetic insights

Alternative smart city agendas

"Cities are for humankind what the telephone booth is for Clark Kent"

Related Podcasts

1

2

3

4

5

6