Deep Learning for Analytical Engine — This repository contains an implementation of a convolutional neural network as a program for Charles Babbage’s Analytical Engine, capable of recognizing handwritten digits to a high degree of accuracy (98.5% if provided with a sufficient amount of training data and left running sufficiently long).
Supervisor Trees in Go — A well-structured Erlang program is broken into multiple independent pieces that communicate via messages, and when a piece crashes, the supervisor of that piece automatically restarts it. […] Even as I have been writing suture, I have on occasion been astonished to flip my screen over to the console of Go program I’ve written with suture, and been surprised to discover that it’s actually been merrily crashing away during my manual testing, but soldiering on so well I didn’t even know.
Developing Quantum Annealer Driven Data Discovery (Joseph Dulny III, Michael Kim) — In this paper, we gain novel insights into the application of quantum annealing (QA) to machine learning (ML) through experiments in natural language processing (NLP), seizure prediction, and linear separability testing.
The Mortality of Companies — Geoffrey West paper: we show that the mortality of publicly traded companies manifests an approximately constant hazard rate over long periods of observation. This regularity indicates that mortality rates are independent of a company’s age. We show that the typical half-life of a publicly traded company is about a decade, regardless of business sector.
Gizmo — a microservice toolkit in Golang from NYT. (via InfoQ)
Intellectual Need and Problem-Free Activity in the Mathematics Classroom (PDF) — Although this is not an empirical study, we use data from observed high school algebra classrooms to illustrate four categories of activity students engage in while feeling little or no intellectual need. We present multiple examples for each category in order to draw out different nuances of the activity, and we contrast the observed situations with ones that would provide various types of intellectual need. Finally, we offer general suggestions for teaching with intellectual need.
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.
uitable — cute library for tabular data in console golang programs.
Did Carnegie Mellon Attack Tor for the FBI? (Bruce Schneier) — The behavior of the researchers is reprehensible, but the real issue is that CERT Coordination Center (CERT/CC) has lost its credibility as an honest broker. The researchers discovered this vulnerability and submitted it to CERT. Neither the researchers nor CERT disclosed this vulnerability to the Tor Project. Instead, the researchers apparently used this vulnerability to deanonymize a large number of hidden service visitors and provide the information to the FBI. Does anyone still trust CERT to behave in the Internet’s best interests? Analogous to the CIA organizing a fake vaccination drive to get close to Osama. “Intelligence” agencies.
The Hit Charade (MIT TR) — Spotify’s deep-learning system still has to be trained using millions of example songs, and it would be perplexed by a bold new style of music. What’s more, such algorithms cannot arrange songs in a creative way. Nor can they distinguish between a truly original piece and yet another me-too imitation of a popular sound. Johnson acknowledges this limitation, and he says human expertise will remain a key part of Spotify’s algorithms for the foreseeable future.
The Future of War is the Distant Past (John Birmingham) — the Naval Academy is hedging against the future by creating cybersecurity midshipmen, and by requiring every midshipman to learn how to do celestial navigation.
What Happens Next Will Amaze You (Maciej Ceglowski) — the next in Maciej’s amazing series of keynotes, where he’s building a convincing case for fixing the Web.
Data Analytics in Sports — O’Reilly research report (free). When it comes to processing stats, competing companies Opta and ProZone use a combination of recording technology and human analysts who tag “events” within the game (much like Vantage Sports). Opta calculates that it tags between 1,600 and 2,000 events per football game — all delivered live.
Tinder and Hook-Up Culture (Vanity Fair) — “There have been two major transitions” in heterosexual mating “in the last four million years,” he says. “The first was around 10,000 to 15,000 years ago, in the agricultural revolution, when we became less migratory and more settled,” leading to the establishment of marriage as a cultural contract. “And the second major transition is with the rise of the Internet.”
Networks Increasing Ad Stuffing — TV audiences (as determined by Nielsen C3 measurements: TV watched both live and three days after the show was first aired on catch-up services) are down 9% year on year, yet ad loads on some networks are up as much as 10% on last year. The dinosaurs are hungry.
Netflix and the Conservation of Attractive Profits (Stratechery) — Note the common element to all three of these companies: all have managed to modularize the production/delivery of their service which has allowed them to move closer to the customer. To put it another way, all of this new value is being created by specialized CRM companies: Airbnb for travelers, Uber for commuters, and Netflix for the bored.
Go as Open Source — keynote from this year’s Gophercon. I’ve been pondering lately how successful open source projects go beyond “anyone can scratch their itch,” and instead actively manage the tendency for scope creep. We’d rather have a small number of features that enable, say, 90% of the target use cases, and avoid the orders of magnitude more features necessary to reach 99% or 100%.
The Universal Data Structure — Given the abysmal state of of today’s software engineering, I believe that a full embrace of the universal hash will result in better, simpler programs. Your weekly dose of snark.
GM: That Car You Bought, We’re Really the Ones Who Own It — GM’s claim is all about copyright and software code, and it’s the same claim John Deere is making about their tractors. The TL;DR version of the argument goes something like this: cars work because software tells all the parts how to operate; the software that tells all the parts to operate is customized code; that code is subject to copyright; GM owns the copyright on that code and that software; a modern car cannot run without that software; it is integral to all systems; therefore, the purchase or use of that car is a licensing agreement; and since it is subject to a licensing agreement, GM is the owner and can allow/disallow certain uses or access. In the future, manufacturers own the secondary market.
Architectural Robots (Robohub) — The concept is named ‘Minibuilders.’ This is a group of robots each performing a specific task. The first robot layers a 15 cm (6 in) footprint or foundation, while a second and a third robot print the rest of the building by climbing over the structures they already printed and laying more material over them. This design is only possible at construction scale where printed layers are solid enough to support a robotic print head.
gigo — Fetching packages in golang can be difficult, especially when juggling multiple packages and private repositories. GIGO (Gigo Installer for Go) tries to solve that problem, effectively being the golang equivalent of Python’s pip.
On Font Design (Kris Sowersby) — The many pairs of hands and eyes involved have made this typeface special for me. For the first time I don’t feel I have ownership of a typeface I have ostensibly “created.” Lovely to read about the design journey for a font.
Why We Use Go — We use Go because it’s boring. Previously, we worked almost exclusively with Python, and after a certain point, it becomes a nightmare. You can bend Python to your will. You can hack it, you can monkey patch it, and you can write remarkably expressive, terse code. It’s also remarkably difficult to maintain and slow.
Unfortunately We Have Renewed Our ICANN Accreditation — You can thank ICANN for this policy, because if it were up to us, and you tasked us with coming up with the most idiotic, damaging, phish-friendly, disaster-prone policy that accomplishes less than nothing and is utterly pointless, I question whether we would have been able to pull it off at this level. We’re simply out of our league here.