Four short links: 29 April 2016

Four short links: 29 April 2016

AI Gym, Open Sourced Xamarin, History Mining, and Club Economics

  1. OpenAI GymA toolkit for developing and comparing reinforcement learning algorithms. It supports teaching agents everything from walking to playing games like Pong or Go.
  2. Open Xamarin — the Xamarin tools are now all MIT-licensed, after the Microsoft acquisition. And show no signs of being abandoned.
  3. Text-Mining the History of MedicineIn this article, we present our efforts to overcome the various challenges faced in the semantic analysis of published historical medical text dating back to the mid 19th century.
  4. A Journal Is a Club (Cameron Neylon) — I’ve been frustrated for a long time with traditional economic analyses of scholarly publishing. They don’t seem to explain what actually happens, and fail to capture critical aspects of what is going on. The lens of club economics seems like it might help to capture more of the reality of what is going on.



Simon King on design at IDEO

The O’Reilly Design Podcast: Team dynamics and culture at IDEO, design education, and design’s next big challenge.

Subscribe to the O’Reilly Design Podcast, our podcast exploring how experience design—and experience designers—are shaping business, the Internet of Things, and other domains. Find us on Stitcher, TuneIn, iTunes, SoundCloud, RSS.


In this week’s Design Podcast episode, I sit down with Simon King, director of Carnegie Mellon University’s Design Center. King is the author of Understanding Industrial Design. We talk about team dynamics and culture at IDEO, extending design education to non-designers, and design’s next big challenge.

Here are a few highlights from our conversation:

CMU’s Design Center and the designer shortage

The idea is behind the Design Center is to bring design to more students across the whole campus and to create new kinds of research opportunities to work with organizations from outside of the university as well. There’s really two levels to think about the designer shortage, and one is what I’m trying to do with the Design Center: increasing the overlap with design that many people might have, where there’s this idea that you’re able to hire someone who’s design savvy, or design literate, or can bring some of the tools of design to their core discipline. That helps, but the bottleneck of ‘there just aren’t enough really skilled designers coming out of master’s programs and that are true experts’—I don’t have a good solution to that. I look at it and I think, ‘design is something you have to learn by doing. You have to build up this set of experiences over time, and it just takes time.’ There’s no silver bullet in terms of meeting the increasing demand.     

Combining it with formal education might be more about a bigger mindset shift and not as much about specific skill building. Another approach is something we’re seeing at CMU, where there’s now a variety of programs. At the master’s level in the School of Design, there’s now options for students that are one year in length, or two years in length, or if they didn’t come in with a design background previously, potentially up to three years in length. It used to be there was just a two-year master’s program. I think that universities are recognizing the need for flexibility there, so that, depending on people’s goals or the time commitment that they’re willing to put in, there’s options for them.              

Read more…


Peter Hoddie on JavaScript for embedded systems

The O’Reilly Hardware Podcast: Hardware abstraction, scripting languages, and user experience.

Subscribe to the O’Reilly Hardware Podcast for insight and analysis about the Internet of Things and the worlds of hardware, software, and manufacturing. Find us on TuneIn, Stitcher, iTunes, SoundCloud, RSS.


This will be a breakout year for JavaScript on embedded systems. Our guest on the Hardware Podcast this week is Peter Hoddie, who founded Kinoma, which makes software and hardware for building JavaScript-powered prototypes. Previously he was one of the original developers of QuickTime at Apple.

Computing power has become so inexpensive, and the JS developer community so large, that the accessibility and fast development times of JavaScript will outweigh the efficiency advantages of C and assembly for all but fairly specialized projects.

Discussion points:

  • Scratch, which begat Blockly, which begat KinomaJS Blocks, which can be used to program the Create embedded platform.
  • Transitioning from software development to hardware development.
  • The importance of developing hardware that is flexible, modular and abstracted. “Why should you throw out your hardware because the software is obsolete?” Hoddie asks.
  • Tools: Hoddie swears by profilers, including Apple’s Instruments and the Kinoma Studio profiler
  • Writing your own JavaScript engine—in this case, Kinoma’s XS6

Read more…

Comments: 6

Using Apache Spark to predict attack vectors among billions of users and trillions of events

The O’Reilly Data Show podcast: Fang Yu on data science in security, unsupervised learning, and Apache Spark.

Subscribe to the O’Reilly Data Show Podcast to explore the opportunities and techniques driving big data and data science: Stitcher, TuneIn, iTunes, SoundCloud, RSS.


In this episode of the O’Reilly Data Show, I spoke with Fang Yu, co-founder and CTO of DataVisor. We discussed her days as a researcher at Microsoft, the application of data science and distributed computing to security, and hiring and training data scientists and engineers for the security domain.

DataVisor is a startup that uses data science and big data to detect fraud and malicious users across many different application domains in the U.S. and China. Founded by security researchers from Microsoft, the startup has developed large-scale unsupervised algorithms on top of Apache Spark, to (as Yu notes in our chat) “predict attack vectors early among billions of users and trillions of events.”

Several years ago, I found myself immersed in the security space and at that time tools that employed machine learning and big data were still rare. More recently, with the rise of tools like Apache Spark and Apache Kafka, I’m starting to come across many more security professionals who incorporate large-scale machine learning and distributed systems into their software platforms and consulting practices.

Read more…


Rob Coneybeer on Nest and the next big thing in hardware

The O’Reilly Hardware Podcast: Virtual reality, robotics, and today’s hardware landscape.

Subscribe to the O’Reilly Hardware Podcast for insight and analysis about the Internet of Things and the worlds of hardware, software, and manufacturing: TuneIn, Stitcher, iTunes, SoundCloud, RSS.


In this new episode of the Hardware Podcast, David Cranor and I talk with Rob Coneybeer, managing director and co-founder of Shasta Ventures, one of the critical first investors in hardware startups including Nest, Fetch Robotics, and Turo (formerly RelayRides).

Discussion points:

  • Why Nest looked like an appealing investment back in 2010
  • Coneybeer’s focus on virtual reality and robotics as the next big things for hardware startups.
  • Why it’s essential for hardware startups to have a long-term plan for improving products after they’re in place, and the importance of over-the-air software updates.
  • The consumer psychology of selling a compelling hardware product, and when to aim for high price and high value. “People are willing to spend money when there’s something that’s really revolutionary,” says Coneybeer.
  • The current state of venture capital investments in hardware startups. While raising later rounds is becoming more difficult, Coneybeer says: “the most interesting, innovative hardware companies will always find capital.”

Read more…


Scott Hurff on designing at Tinder

The O’Reilly Design Podcast: Design at Tinder, Awkward UI, and the UI Stack.

Subscribe to the O’Reilly Design Podcast, our podcast exploring how experience design—and experience designers—are shaping business, the Internet of Things, and other domains: Stitcher, TuneIn, iTunes, SoundCloud, RSS.


In this week’s Design Podcast episode, I sit down with Scott Hurff, product manager and lead designer at Tinder, Inc. Hurff is the author of Designing Products People Love. In this episode, we talk about how Tinder approaches design, avoiding awkward UI, and why customer research is the most important skill for future designers.

Here are a few highlights from our conversation:

Questions of structure

At Tinder, the product team is about five people, six people. What’s interesting is that we’re trying to grow really quickly. There’s a give and take on how we divide up product design responsibilities and product management responsibilities. There is a lot of engineering talent here, and they need a lot of product to work on. It’s a matter of, how do we structure ourselves so we can give them thought-through, packaged-up, ready-to-go ideas and concepts while still hammering out the details in time.

Design as a full-contact sport

Design is such a part of the Tinder experience. It may not seem like that’s the case because it’s such a simple app, but that’s only because everything goes through this distillation process. You have to really fight for real estate and your idea. Design’s really a full-contact sport here. You have to bring in all the big guns to make your case. Sometimes these can be really long debates, but they’re good; they’re healthy. They get the ideas out on the table, and a lot of times, design really has to be put through its bases to prove itself.

Read more…