Skip to content

Justin's Linklog Posts

Links for 2015-06-23

Links for 2015-06-22

Links for 2015-06-21

  • jwz on Inceptionism

    “Shoggoth ovipositors”:

    So then they reach inside to one of the layers and spin the knob randomly to fuck it up. Lower layers are edges and curves. Higher layers are faces, eyes and shoggoth ovipositors. [….] But the best part is not when they just glitch an image — which is a fun kind of embossing at one end, and the “extra eyes” filter at the other — but is when they take a net trained on some particular set of objects and feed it static, then zoom in, and feed the output back in repeatedly. That’s when you converge upon the platonic ideal of those objects, which — it turns out — tend to be Giger nightmare landscapes. Who knew. (I knew.)
    This stuff is still boggling my mind. All those doggy faces! That is one dog-obsessed ANN.

    (tags: neural-networks ai jwz funny shoggoths image-recognition hr-giger art inceptionism)

Links for 2015-06-19

Links for 2015-06-18

  • Inceptionism: Going Deeper into Neural Networks

    This is amazing, and a little scary.

    If we choose higher-level layers, which identify more sophisticated features in images, complex features or even whole objects tend to emerge. Again, we just start with an existing image and give it to our neural net. We ask the network: “Whatever you see there, I want more of it!” This creates a feedback loop: if a cloud looks a little bit like a bird, the network will make it look more like a bird. This in turn will make the network recognize the bird even more strongly on the next pass and so forth, until a highly detailed bird appears, seemingly out of nowhere.
    An enlightening comment from the G+ thread:
    This is the most fun we’ve had in the office in a while. We’ve even made some of those ‘Inceptionistic’ art pieces into giant posters. Beyond the eye candy, there is actually something deeply interesting in this line of work: neural networks have a bad reputation for being strange black boxes that that are opaque to inspection. I have never understood those charges: any other model (GMM, SVM, Random Forests) of any sufficient complexity for a real task is completely opaque for very fundamental reasons: their non-linear structure makes it hard to project back the function they represent into their input space and make sense of it. Not so with backprop, as this blog post shows eloquently: you can query the model and ask what it believes it is seeing or ‘wants’ to see simply by following gradients. This ‘guided hallucination’ technique is very powerful and the gorgeous visualizations it generates are very evocative of what’s really going on in the network.?

    (tags: art machine-learning algorithm inceptionism research google neural-networks learning dreams feedback graphics)

Links for 2015-06-17

Links for 2015-06-15

  • How We Moved Our API From Ruby to Go and Saved Our Sanity

    Parse on their ditching-Rails story. I haven’t heard a nice thing about Ruby or Rails as an operational, production-quality platform in a long time :(

    (tags: go ruby rails ops parse languages platforms)

  • VPC Flow Logs

    we are introducing Flow Logs for the Amazon Virtual Private Cloud.  Once enabled for a particular VPC, VPC subnet, or Elastic Network Interface (ENI), relevant network traffic will be logged to CloudWatch Logs for storage and analysis by your own applications or third-party tools. You can create alarms that will fire if certain types of traffic are detected; you can also create metrics to help you to identify trends and patterns. The information captured includes information about allowed and denied traffic (based on security group and network ACL rules). It also includes source and destination IP addresses, ports, the IANA protocol number, packet and byte counts, a time interval during which the flow was observed, and an action (ACCEPT or REJECT).

    (tags: ec2 aws vpc logging tracing ops flow-logs network tcpdump packets packet-capture)

  • Tim Hunt “jokes” about women scientists. Or not. (with image, tweets) · deborahblum · Storify

    ‘[Tim Hunt] said that while he meant to be ironic, he did think it was hard to collaborate with women because they are too emotional – that he was trying to be honest about the problems.’ So much for the “nasty twitter took my jokes seriously” claims then.

    (tags: twitter science misogyny women tim-hunt deborah-blum journalism)

  • Why I dislike systemd

    Good post, and hard to disagree.

    One of the “features” of systemd is that it allows you to boot a system without needing a shell at all. This seems like such a senseless manoeuvre that I can’t help but think of it as a knee-jerk reaction to the perception of Too Much Shell in sysv init scripts. In exactly which universe is it reasonable to assume that you have a running D-Bus service (or kdbus) and a filesystem containing unit files, all the binaries they refer to, all the libraries they link against, and all the configuration files any of them reference, but that you lack that most ubiquitous of UNIX binaries, /bin/sh?

    (tags: history linux unix systemd bsd system-v init ops dbus)

  • Adrian Colyer reviews the Twitter Heron paper

    ouch, really sounds like Storm didn’t cut the muster. ‘It’s hard to imagine something more damaging to Apache Storm than this. Having read it through, I’m left with the impression that the paper might as well have been titled “Why Storm Sucks”, which coming from Twitter themselves is quite a statement.’ If I was to summarise the lessons learned, it sounds like: backpressure is required; and multi-tenant architectures suck.

    (tags: storm twitter heron big-data streaming realtime backpressure)

Links for 2015-06-14

  • Security theatre at Allied Irish Banks

    Allied Irish Banks’s web and mobile banking portals are ludicrously insecure. Vast numbers of accounts have easily-guessable registration numbers and are thus ‘protected’ by a level of security that is twice as easy to crack as would be provided by a single password containing only two lowercase letters. A person of malicious intent could easily gain access to hundreds, possibly thousands, of accounts as well as completely overwhelm the branch network by locking an estimated several 100,000s of people out of their online banking. Both AIB and the Irish Financial Services Ombudsman have refused to respond meaningfully to multiple communications each in which these concerns were raised privately.

    (tags: aib banking security ireland hacking ifso online-banking)

  • Leveraging AWS to Build a Scalable Data Pipeline

    Nice detailed description of an auto-scaled SQS worker pool

    (tags: sqs aws ec2 auto-scaling asg worker-pools architecture scalability)

Links for 2015-06-13

  • China’s Spies Hit the Blackmail Jackpot With Data on 4 Million Federal Workers

    The Daily Beast is scathing re the OPM hack:

    Here’s where things start to get scary. Whoever has OPM’s records knows an astonishing amount about millions of federal workers, members of the military, and security clearance holders. They can now target those Americans for recruitment or influence. After all, they know their vices, every last one—the gambling habit, the inability to pay bills on time, the spats with former spouses, the taste for something sexual on the side—since all that is recorded in security clearance paperwork. (To get an idea of how detailed this gets, you can see the form, called an SF86, here.) Speaking as a former counterintelligence officer, it really doesn’t get much worse than this.

    (tags: daily-beast sf86 clearance us-government america china cyberwar hacking opm privacy)

Links for 2015-06-11

  • Facebook Infer

    New static analysis goodnews, freshly open-sourced by Facebook:

    Facebook Infer uses logic to do reasoning about a program’s execution, but reasoning at this scale — for large applications built from millions of lines of source code — is hard. Theoretically, the number of possibilities that need to be checked is more than the number of estimated atoms in the observable universe. Furthermore, at Facebook our code is not a fixed artifact but an evolving system, updated frequently and concurrently by many developers. It is not unusual to see more than a thousand modifications to our mobile code submitted for review in a given day. The requirements on the program analyzer then become even more challenging because we expect a tool to report quickly on these code modifications — in the region of 10 minutes — to fit in with developers’ workflow. Coping with this scale and velocity requires advanced mathematical techniques. Facebook Infer uses two such techniques: separation logic and bi-abduction. Separation logic is a theory that allows Facebook Infer’s analysis to reason about small, independent parts of the application storage, rather than having to consider the entirety of the memory potentially at every step. That would be a daunting task on modern processors with their large addressable virtual memories. Bi-abduction is a logical inference technique that allows Facebook Infer to discover properties about the behavior of independent parts of the application code. By storing these properties between runs, Facebook Infer needs to analyze only the parts of the software that have changed, reusing the results of its previous analysis where it can. By combining these approaches, our analyzer is able to find complex problems in modifications to an application built from millions of lines of code, in minutes.
    (via Bryan O’Sullivan)

    (tags: via:bos infer facebook static-analysis lint code java ios android coding bugs)

  • The Tamborzão Goes to Thailand

    This is great. the story of how cheesy funk carioca tune “A Minha Amiga Fran” managed to become “Kawo Kawo” and become a massive hit in Thailand

    (tags: thai brazil carioca music dance-music kawo-kawo)

Links for 2015-06-10

Links for 2015-06-08

Links for 2015-06-04

Links for 2015-06-03

  • Performance Testing at LMAX

    Good series of blog posts on the LMAX trading platform’s performance testing strategy — they capture live traffic off the wire, then build statistical models simulating its features. See also http://epickrram.blogspot.co.uk/2014/07/performance-testing-at-lmax-part-two.html and http://epickrram.blogspot.co.uk/2014/08/performance-testing-at-lmax-part-three.html .

    (tags: performance testing tests simulation latency lmax trading sniffing packet-capture)

  • The Violence of Algorithms: Why Big Data Is Only as Smart as Those Who Generate It

    The modern state system is built on a bargain between governments and citizens. States provide collective social goods, and in turn, via a system of norms, institutions, regulations, and ethics to hold this power accountable, citizens give states legitimacy. This bargain created order and stability out of what was an increasingly chaotic global system. If algorithms represent a new ungoverned space, a hidden and potentially ever-evolving unknowable public good, then they are an affront to our democratic system, one that requires transparency and accountability in order to function. A node of power that exists outside of these bounds is a threat to the notion of collective governance itself. This, at its core, is a profoundly undemocratic notion—one that states will have to engage with seriously if they are going to remain relevant and legitimate to their digital citizenry who give them their power.

    (tags: palantir algorithms big-data government democracy transparency accountability analytics surveillance war privacy protest rights)

Links for 2015-06-02

Links for 2015-05-29

Links for 2015-05-28

  • I Fooled Millions Into Thinking Chocolate Helps Weight Loss

    “Slim by Chocolate!” the headlines blared. A team of German researchers had found that people on a low-carb diet lost weight 10 percent faster if they ate a chocolate bar every day. It made the front page of Bild, Europe’s largest daily newspaper, just beneath their update about the Germanwings crash. From there, it ricocheted around the internet and beyond, making news in more than 20 countries and half a dozen languages. It was discussed on television news shows. It appeared in glossy print, most recently in the June issue of Shape magazine (“Why You Must Eat Chocolate Daily”, page 128). Not only does chocolate accelerate weight loss, the study found, but it leads to healthier cholesterol levels and overall increased well-being. The Bild story quotes the study’s lead author, Johannes Bohannon, Ph.D., research director of the Institute of Diet and Health: “The best part is you can buy chocolate everywhere.” I am Johannes Bohannon, Ph.D. Well, actually my name is John, and I’m a journalist. I do have a Ph.D., but it’s in the molecular biology of bacteria, not humans. The Institute of Diet and Health? That’s nothing more than a website. Other than those fibs, the study was 100 percent authentic. My colleagues and I recruited actual human subjects in Germany. We ran an actual clinical trial, with subjects randomly assigned to different diet regimes. And the statistically significant benefits of chocolate that we reported are based on the actual data. It was, in fact, a fairly typical study for the field of diet research. Which is to say: It was terrible science. The results are meaningless, and the health claims that the media blasted out to millions of people around the world are utterly unfounded.
    Interesting bit: the online commenters commenting on the published stories quickly saw through the bullshit. Why can’t the churnalising journos do that?

    (tags: chocolate journalism science diet food churnalism pr bild health clinical-trials papers peer-review research)

  • Snake-Oil Superfoods

    mainly interesting for the dataviz and the Google-Doc-driven backend. wish they published the script though

    (tags: google snake-oil superfoods food dataviz bubble-race-chart graphics infographics google-docs spreadsheets)

Links for 2015-05-27

  • Three Questions to Answer When Reporting an Error

    Very long, but tl;dr:

    the trick to creating an effective error message is to answer the 3 Questions within your message: What is the error? What was the probable cause of the error? What is the probable remedy?

    (tags: errors ui ux reporting logging coding)

  • Volvo says horrible ‘self-parking car accident’ happened because driver didn’t have ‘pedestrian detection’

    Grim meathook future, courtesy of Volvo:

    “The Volvo XC60 comes with City Safety as a standard feature however this does not include the Pedestrian detection functionality […] The pedestrian detection feature […] costs approximately $3,000.
    However, there’s another lesson here, in crappy car UX and the risks thereof:
    But even if it did have the feature, Larsson says the driver would have interfered with it by the way they were driving and “accelerating heavily towards the people in the video.” “The pedestrian detection would likely have been inactivated due to the driver inactivating it by intentionally and actively accelerating,” said Larsson. “Hence, the auto braking function is overrided by the driver and deactivated.” Meanwhile, the people in the video seem to ignore their instincts and trust that the car assumed to be endowed with artificial intelligence knows not to hurt them. It is a sign of our incredible faith in the power of technology, but also, it’s a reminder that companies making AI-assisted vehicles need to make safety features standard and communicate clearly when they aren’t.

    (tags: self-driving-cars cars ai pedestrian computer-vision volvo fail accidents grim-meathook-future)

  • iPhone UTF-8 text vulnerability

    ‘Due to how the banner notifications process the Unicode text. The banner briefly attempts to present the incoming text and then “gives up” thus the crash’. Apparently the entire Springboard launcher crashes.

    (tags: apple vulnerability iphone utf-8 unicode fail bugs springboard ios via:abetson)

Links for 2015-05-26

  • Schedule Recurring AWS Lambda Invocations With The Unreliable Town Clock (UTC)

    The Unreliable Town Clock (UTC) is a new, free, public SNS Topic (Amazon Simple Notification Service) that broadcasts a “chime” message every quarter hour to all subscribers. It can send the chimes to AWS Lambda functions, SQS queues, and email addresses. You can use the chime attributes to run your code every fifteen minutes, or only run your code once an hour (e.g., when minute == “00”) or once a day (e.g., when hour == “00” and minute == “00”) or any other series of intervals. You can even subscribe a function you only want to run only once at a specific time in the future: Have the function ignore all invocations until it’s after the time it wants. When it is time, it can perform its job, then unsubscribe itself from the SNS Topic.

    (tags: alestic aws lambda cron time clock periodic-tasks recurrence hacks)

Links for 2015-05-25

Links for 2015-05-18

Links for 2015-05-17

  • ‘Can People Distinguish Pâté from Dog Food?’

    Ugh.

    Considering the similarity of its ingredients, canned dog food could be a suitable and inexpensive substitute for pâté or processed blended meat products such as Spam or liverwurst. However, the social stigma associated with the human consumption of pet food makes an unbiased comparison challenging. To prevent bias, Newman’s Own dog food was prepared with a food processor to have the texture and appearance of a liver mousse. In a double-blind test, subjects were presented with five unlabeled blended meat products, one of which was the prepared dog food. After ranking the samples on the basis of taste, subjects were challenged to identify which of the five was dog food. Although 72% of subjects ranked the dog food as the worst of the five samples in terms of taste (Newell and MacFarlane multiple comparison, P<0.05), subjects were not better than random at correctly identifying the dog food.

    (tags: pate food omgwtf science research dog-food meat economics taste flavour)

  • Redditor runs the secret Python code in Ex Machina

    and finds:

    when you run with python2.7 you get the following: ISBN = 9780199226559 Which is Embodiment and the inner life: Cognition and Consciousness in the Space of Possible Minds. and so now I have a lot more respect for the Director.

    (tags: python movies ex-machina cool books easter-eggs)

  • Metalwoman beer recipe

    via the Dublin Ladies Beer Society ;)

    (tags: metalman metalwoman recipes beer brewing hops dlbs)

Links for 2015-05-14

Links for 2015-05-12

  • Input: Fonts for Code

    Non-monospaced coding fonts! I’m all in favour…

    As writing and managing code becomes more complex, today’s sophisticated coding environments are evolving to include everything from breakpoint markers to code folding and syntax highlighting. The typography of code should evolve as well, to explore possibilities beyond one font style, one size, and one character width.

    (tags: input fonts via:its typography code coding font text ide monospace)

  • Apache HTrace

    a Zipkin-compatible distributed-system tracing framework in Java, in the Apache Incubator

    (tags: zipkin tracing trace apache incubator java debugging)

  • Intel speeds up etcd throughput using ADR Xeon-only hardware feature

    To reduce the latency impact of storing to disk, Weaver’s team looked to buffering as a means to absorb the writes and sync them to disk periodically, rather than for each entry. Tradeoffs? They knew memory buffers would help, but there would be potential difficulties with smaller clusters if they violated the stable storage requirement. Instead, they turned to Intel’s silicon architects about features available in the Xeon line. After describing the core problem, they found out this had been solved in other areas with ADR. After some work to prove out a Linux OS supported use for this, they were confident they had a best-of-both-worlds angle. And it worked. As Weaver detailed in his CoreOS Fest discussion, the response time proved stable. ADR can grab a section of memory, persist it to disk and power it back. It can return entries back to disk and restore back to the buffer. ADR provides the ability to make small (<100MB) segments of memory “stable” enough for Raft log entries. It means it does not need battery-backed memory. It can be orchestrated using Linux or Windows OS libraries. ADR allows the capability to define target memory and determine where to recover. It can also be exposed directly into libs for runtimes like Golang. And it uses silicon features that are accessible on current Intel servers.

    (tags: kubernetes coreos adr performance intel raft etcd hardware linux persistence disk storage xeon)

Links for 2015-05-11

  • streamtools: a graphical tool for working with streams of data | nytlabs

    Visual programming, Yahoo! Pipes style, back again:

    we have created streamtools – a new, open source project by The New York Times R&D Lab which provides a general purpose, graphical tool for dealing with streams of data. It provides a vocabulary of operations that can be connected together to create live data processing systems without the need for programming or complicated infrastructure. These systems are assembled using a visual interface that affords both immediate understanding and live manipulation of the system.
    via Aman

    (tags: via:akohli streaming data nytimes visual-programming coding)

  • MappedBus

    a Java based low latency, high throughput message bus, built on top of a memory mapped file; inspired by Java Chronicle with the main difference that it’s designed to efficiently support multiple writers – enabling use cases where the order of messages produced by multiple processes are important. MappedBus can be also described as an efficient IPC mechanism which enable several Java programs to communicate by exchanging messages.

    (tags: ipc java jvm mappedbus low-latency mmap message-bus data-structures queue message-passing)

Links for 2015-05-10

  • Amazon’s Drone Delivery Patent Just Feels Like Trolling At This Point

    Oh dear, Amazon.

    These aren’t actual technologies yet. […] All of which underscores that Amazon might never ever ever ever actually implement delivery drones. The patent paperwork was filed nearly a year after Amazon’s splashy drone program reveal on 60 Minutes. At the time we called it revolutionary marketing because, you know, delivery drones are technical and logistical madness, not to mention that commercial drone use is illegal right now. Although, in fairness the FAA did just relax some rules so that Amazon could test drones. At this point it feels like Amazon is just trolling. It’s trolling us with public relations BS about its future drones, and it’s trolling future competitors — Google is also apparently working on this — so that if somebody ever somehow does anything relating to drone delivery, Amazon can sue them. If I’m wrong, I’ll deliver my apology via Airmail.

    (tags: amazon trolling patents uspto delivery drones uavs competition faa)

  • Red Hat on rkt vs Docker

    This is like watching a train-wreck in slow motion on Groundhog Day. We, in the broader Linux and open source community, have been down this path multiple times over the past fifteen years, specifically with package formats. While there needs to be room for experimentation, having two incompatible specs driven by two startups trying to differentiate and in direct competition is *not* a good thing. It would be better for the community and for everyone who depends on our collective efforts if CoreOS and Docker collaborated on a standardized common spec, image format, and distribution protocol. To this end, we at Red Hat will continue to contribute to both initiatives with the goal of driving convergence.

    (tags: rkt docker appc coreos red-hat dpkg rpm linux packaging collaboration open-source)

Links for 2015-05-09

Links for 2015-05-08

Links for 2015-05-07

  • KillBiller

    Excellent mobile-phone plan comparison site for the Irish market, using apps which you install and which analyse your call history, data usage, etc. over the past month to compute the optimal plan based on your usage. Pretty amazing results in my case! The only downside is the privacy policy, which allows the company to resell your usage data (anonymised, and in aggregate) — I’d really prefer if this wasn’t the case :(

    (tags: mobile-phones shopping tesco emobile 3g 4g ireland plans comparison-shopping killbiller via:its)

  • Family in No poster Says YES to Marriage Equality | Amnesty International

    Beyond the politics, the risks of stock photo usage are pretty evident too:

    “In 2014, as a young family, we did a photo shoot with a photographer friend to get some nice shots for the family album. No money was exchanged – we got nice photos for free, they got nice images for their portfolio. As part of this agreement, we agreed to let them upload them to a stock photo album. We knew that these were available for purchase and we gave permission. Perhaps, naïvely, we imagined that on the off chance that any was ever selected, it might be for a small magazine or website. To confirm, we have not received any money for the photo – then or now, and nor do we expect any. We were surprised and upset to see that the photo was being used as part of a campaign with which we do not agree. We completely support same-sex marriage, and we believe that same-sex couples’ should of course be able to adopt, as we believe that they are equally able to provide children with much-needed love and care. To suggest otherwise is offensive to us, and to many others.”

    (tags: ssm ireland politics amnesty stock-photos ip rights photos campaigns ads)

  • Lambda: Bees with Frickin’ Laser Beams

    a HTTP testing tool in AWS Lambda. nice enough, but still a toy…

    (tags: lambda aws node javascript hacks http load-testing)

Links for 2015-05-06