Deep Visual Analogy-Making (PDF) — In this paper, we develop a novel deep network trained end-to-end to perform visual analogy making, which is the task of transforming a query image according to an example pair of related images. Open source code from the paper also available.
Samsung’s TV and Privacy Gets More Awkward — Samsung has now issued a new statement clarifying how the voice activation feature works. “If a consumer consents and uses the voice recognition feature, voice data is provided to a third party during a requested voice command search,” Samsung said in a statement. “At that time, the voice data is sent to a server, which searches for the requested content then returns the desired content to the TV.” It only seems creepy until you give in and nothing bad happens, then you normalise the creepy.
2015 Robot Numbers (RoboHub) — The Robotic Industries Association (RIA), representing North American robotics, reported […] 2015 set new records and showed a 14% increase in units and 11% in dollars over 2014. The automotive industry was the primary growth sector, with robot orders increasing 19% year over year. Non-automotive robot orders grew at 5%.
Mozilla, Caribou Digital Release Report Exploring the Global App Economy (Mark Surman) — The emerging markets are the 1% — meaning, they earn 1% of total app economy revenue. 95% of the estimated value in the app economy is captured by just 10 countries, and 69% of the value is captured by just the top three countries. Excluding China, the 19 countries considered low- or lower-income accounted for only 1% of total worldwide value. Developers in low-income countries struggle to export to the global stage. About one-third of developers in the sample appeared only in their domestic market.
Spermbots — Researchers from the Institute for Integrative Nanosciences at IFW Dresden in Germany have successfully tested tiny, magnetically-driven power suits for individual sperm that can turn them into steerable cyborg “spermbots” that can be remote controlled all the way to the egg. But can they make an underwire bra that the washing machine doesn’t turn into a medieval torture device?
What’s Eating Silicon Valley — In 2014, more Harvard Business School Grads went into technology than into banking for the first time since the dot-com era. […] another reason Wall Street had trouble maintaining goodwill was because of some of the attributes above—hard-charging, too much too soon, parallel reality, money flowing everywhere, rich white guys, etc. The Wall St comparison was new to me, but I can see it as a goodwill risk.
OpenTrons — $3,000 open source personal lab robot for science, with downloadable/shareable protocols.
Why Big Companies Keep Failing: The Stack Fallacy — you’re more likely to succeed if you expand down (to supplant your suppliers) than up (to build the products that are built on top of your product) because you’re a customer of your suppliers, so you know what good product-market fit will look like, but you’re just fantasizing that you can supplant your downstream value.
Is Caffeine a Cognitive Enhancer? (PDF) — Two general mechanisms may account for most of the observed effects of caffeine on performance: (1) an indirect, non-specific ‘arousal’ or ‘processing resources’ factor, presumably explaining why the effects of caffeine are generally most pronounced when task performance is sustained or degraded under suboptimal conditions; and (2) a more direct and specific ‘perceptual-motor’ speed or efficiency factor that may explain why, under optimal conditions, some aspects of human performance and information processing, in particular those related to sensation, perception, motor preparation, and execution, are more sensitive to caffeine effects than those related to cognition, memory, and learning. See also Smith 2005‘s caffeine led to a more positive mood and improved performance on a number of tasks. Different effects of caffeine were seen depending on the person’s level of arousal. Linear effects of caffeine dose were also observed. This is evidence against the argument that behavioral changes due to caffeine are merely the reversal of negative effects of a long period of caffeine abstinence. (via cogsci.stackexchange.com)
On Stars and Thinking Things Through (Courtney Johnston) — Matt (to my eyes, anyway) doesn’t have a singular ‘thing’: he has this kind of spangly web of interests and skills that coalesces around a line of enquiry and results in the making or doing of a thing, and these things in turn become part of that web and generate further experiments and thinking. Seconded.
Human-like Robot — and just like a real woman, the first paragraphs about the robot focus on soft skin and flowing brunette hair not how well she does her job. Progress!
Bank of America Loading up on Bitcoin Patents — The wide-ranging patents cover everything from a “cryptocurrency transaction payment system” which would let users make transactions using cryptocurrency, to risk detection, storing cryptocurrencies offline, and using the blockchain to measure fraudulent activity.
Narcos GPS-Spoofing Border Drones — not only are the border drones expensive and ineffective, now they’re being tricked. Basic trade-off: more reliability or longer flight times?
A Model Explanation System (PDF) — you can explain any machine-learned decision, though not necessarily the way the model came to the decision. Confused? This summary might help. Explainability is not a property of the model.
Bro — open source intrusion and anomaly detection service, turns everything into events that you can run scripts against. Good pedigree (Vern Paxson, a TCP/IP elder god) despite the wince-inducing name (at least it isn’t “brah”).
Contempt Culture (Aurynn) — for a culture that now prides itself on continuous improvement and blameless post-mortems and so on, we’re blind to a contempt culture that produces cults of criticism like “PHP isn’t a real programming language,” etc., where the targets of the criticism are pathways disproportionately taken by women and minorities. I’m embarrassed by how much of 2001-era Nat I recognise in Aurynn’s description.
Deep Learning Robot — Built for advanced research in robotics and artificial intelligence (deep learning). Pre-installed Google TensorFlow, Robot Operating System (ROS), Caffe, Torch, Theano, CUDA, and cuDNN.
Juniper ScreenOS Backdoor — here’s the ssh password that’ll get you into any unpatched Juniper firewall, courtesy a backdoor that will be keeping network admins and CEOs alike awake and unhappy around the world. The interesting analysis with long-term effects will be “how the hell did it get in there?”
Face Director — Disney software to match faces between takes. We demonstrate that our method can synthesize visually believable performances with applications in emotion transition, performance correction, and timing control.
Move Fast and Fix Things — blow by blow of an engineering rewrite of some key functionality at GitHub, interesting from a “oh so that’s how they do it” point of view (if blow-by-blow engineering rewrites qualify as “interesting” to you).
Old Book Illustrations — public domain book illustrations, tagged and searchable. Yes, like Font Awesome of engraving.
Toxic Workers (PDF) — In comparing the two costs, even if a firm could replace an average worker with one who performs in the top 1%, it would still be better off by replacing a toxic worker with an average worker by more than two-to-one. Harvard Business School research. (via Fortune)
Replacing Sawzall (Google) — At Google, most Sawzall analysis has been replaced by Go […] we’ve developed a set of Go libraries that we call Lingo (for Logs in Go). Lingo includes a table aggregation library that brings the powerful features of Sawzall aggregation tables to Go, using reflection to support user-defined types for table keys and values. It also provides default behavior for setting up and running a MapReduce that reads data from the logs proxy. The result is that Lingo analysis code is often as concise and simple as (and sometimes simpler than) the Sawzall equivalent.
The O-Ring Theory of DevOps (Adrian Colyer) — Small differences in quality (i.e, in how quickly and accurately you perform each stage of your DevOps pipeline) quickly compound to make very large differences between the performance of the best-in-class and the rest.